Solutions to Section 1

Exercise 1.1
Show that |a| a and |a| a.

Solution.
This follows from the fact that max{a, a} a and max{a, a} a

Exercise 1.2
Show that 
a if a 0
|a| =
a if a < 0
That is, the absolute value function is a piecewise defined function. Graph
this function in the rectangular coordinate system.

Solution.
If a 0 then a 0 so that |a| = max{a, a} = a. If a < 0 then a > 0 so
that |a| = max{a, a} = a. The graph is shown in Figure 1.

Figure 1

Exercise 1.3
Show that |a| 0 with |a| = 0 if and only if a = 0.

Solution.
We see from the graph of |a| that |a| 0 and |a| = 0 if and only if a = 0

Exercise 1.4
Show that if |a| = |b| then a = b.

Solution.
Suppose first that a 0. Then |a| = a.
If b 0 then |b| = b. In this case, |a| = |b| implies that a = b.

1
 If b < 0 then |b| = b. In this case, |a| = |b| implies that a = b.
Suppose now that a < 0. Then |a| = a.
If b 0 then |b| = b. In this case, |a| = |b| implies that a = b which is
the same as a = b.
If b < 0 then |b| = b. In this case, |a| = |b| implies that a = b which
is equivalent to a = b

Exercise 1.5
Solve the equation |3x 2| = |5x + 4|.

Solution.
We have either 3x 2 = 5x + 4 or 3x 2 = (5x + 4). Solving the first
equation we find x = 3. Solving the second equation, we find x = 41

Exercise 1.6
Show that | a| = |a|.

Solution.
If a 0 then a < 0. Thus, |a| = a and | a| = (a) = a. Hence,
| a| = |a|. If a < 0 then a > 0. In this case, |a| = a and | a| = a.
That is, | a| = |a|

Exercise 1.7
Show that |ab| = |a| |b|.

Solution.
Suppose first that a 0. Then |a| = a.
If b 0 then |b| = b. Moreover, ab 0. In this case, |ab| = ab = |a| |b|.
If b < 0 then |b| = b. Moreover, ab 0. In this case, |ab| = ab =
a(b) = |a| |b|.
Suppose now that a < 0. Then |a| = a.
If b 0 then |b| = b. Moreover, ab 0. In this case, |ab| = ab = (a)b =
|a| |b|.
If b < 0 then |b| = b. Moreover, ab > 0. In this case, |ab| = ab = |a| |b|

Exercise 1.8
 
Show that  a1  = 1
|a|
, where a 6= 0.

2
Solution.  
If a > 0 then a1 > 0. Thus,  a1  = 1
a
= 1
|a|
. If a < 0 then 1
a
< 0. Thus,
1
  = 1 = 1 = 1
a a a |a|

Exercise 1.9
|a|
 
Show that  ab  = |b|
where b 6= 0.
Solution.  
Using both Exercise 1.7 and Exercise 1.8 we can write the following  ab  =
a  = |a|  1  = |a| 1 = |a|
 1  
b b |b| |b|

Exercise 1.10
Show that for any two real numbers a and b we have ab |a| |b|.
Solution.
From Exercise 1.1, we have ab |ab| = |a| |b|, where we used Exercise 1.7
Exercise 1.11
Recall that a number b 0 is the square root of a number a, written

a = b, if and only if a = b2 . Show that

a2 = |a|.
Solution.
Since (a)2 = a2 we can write a2 = a if a < 0 or a2 = a if a 0. But
this is equivalent to writing a2 = |a| by Exercise 1.2
Exercise 1.12
suppose that A and B are points on a coordinate line that have coordinates
a and b, respectively. Show that |a b| is the distance between the points A
and B. Thus, if b = 0, |a| measures the distance from the number a to the
origin.
Solution.
Let d be the distance between A and B. If a < b (See Figure 2(a)) then
d = b a and |a b| = (a b) = b a = d. If a > b (See Figure 2(b)) then
d = a b and |a b| = a b = d

3
 Figure 2

Exercise 1.13
Graph the portion of the real line given by the inequality
(a) |x a| <
(b) 0 < |x a| <
where > 0. Represent each graph in interval notation.

Solution.
(a) The graph is given in Figure 3(a). In interval notation, we have (a
, a + ).
(b) The graph is given in Figure 3(b). In interval notation, we have (a
, a) (, a + )

Figure 3

Exercise 1.14
Show that |x a| < k if and only if a k < x < a + k, where k > 0.

Solution.
From the previous exercise, we can write |x a| < k a k < x < a + k

Exercise 1.15
Show that |x a| > k if and only if x < a k or x > a + k, where k 0.

Solution.
Using Figure 4, we see that |x a| > k x a < k or x a > k. This is
equivalent to x < a k or x > a + k

Figure 4

4
Exercise 1.16
Solve each of the following inequalities: (a) |2x 3| < 5 and (b) |x + 4| > 2.

Solution.
(a) Using Exercise 1.14 we can write |2x 3| < 5 5 < 2x 3 < 5
5 + 3 < 2x < 5 + 3 1 < x < 4.
(b) Using Exercise 1.15 we can write |x + 4| > 2 x + 4 < 2 or x + 4 > 2
which is equivalent to x < 6 or x > 2

Exercise 1.17 (Triangle inequality)

Use Exercise 1.1, Exercise 1.7, and the expansion of (|a + b|)2 to establish
the inequality
|a + b| |a| + |b|,
where a and b are arbitrary real numbers.

Solution.
We have

(|a + b|)2 = (a + b)(a + b) = a2 + 2ab + b2 a2 + 2|a| |b| + b2 = (|a| + |b|)2 .

Now, the result follows by taking the square root of both sides

Exercise 1.18
Show that for any real numbers a and b we have |a| |b| |a b|. Hint:
Notice that a = (a b) + b.

Solution.
Using Exercise 1.17 we can write |a| = |(a b) + b| |a b| + |b|. Subtracting
|b| from both sides to obtain the desired inequaltiy

Exercise 1.19
Let a R. Show that max{a, 0} = 12 (a + |a|) and min{a, 0} = 21 (a |a|).

Solution.
The results are clear if a = 0. If a > 0 then |a| = a and max{a, 0} = a =
1
2
(a + |a|). If a < 0 then |a| = a max{a, 0} = 0 = 21 (a + |a|). Likewise for
the minimum

Exercise 1.20
Show that |a + b| = |a| + |b| if and only if ab 0.

5
Solution.
Suppose first that |a + b| = |a| + |b|. Squaring both sides we find a2 + 2ab +
b2 = a2 + 2|a||b| + b2 or equivalently ab = |a||b| = |ab|. But this is true
only when ab 0. Conversely, suppose that ab 0. If a = 0 we have
|0 + b| = |b| = |0| + |b|. Likewise when b = 0. So assume that ab > 0. Suppose
that a > 0 and b > 0. Then a+b > 0 and in this case |a+b| = a+b = |a|+|b|.
Similar argument when a < 0 and b < 0

Exercise 1.21
Suppose 0 < x < 12 . Simplify x+3
|2x2 +5x3|
.

Solution.
We have
x+3 x+3 1 1
= = =
|2x2 + 5x 3| |(2x 1)(x + 3)| |2x 1| 1 2x

Exercise 1.22
Write the function f (x) = |x + 2| + |x 4| as a piecewise defined function.
Sketch its graph.

Solution.
We have 
x+2 if x 2
|x + 2| =
(x + 2) if x < 2
Likewise, 
x4 if x 4
|x 4| =
(x 4) if x < 4
Combining we find

2x + 2 if x < 2
f (x) = 6 if 2 x < 4
2x if x 4

The graph is given below

6
Exercise 1.23
Prove that ||a| |b|| |a b| for any real numbers a and b.

Solution.
We have |b| |a| |b a| = |a b| so that |a b| |a| |b| |a b by
Exercise 1.18. Now using Exercise 1.14, the result follows

Exercise 1.24
Solve the equation 4|x 3|2 3|x 3| = 1.

Solution.
Let u = |x 3|. Then 4u2 3u 1 = 0 implies u = 1 or u = 1
4
. Since u 0
we must have |x 3| = u = 1. Hence, x 3 = 1 or x 3 = 1. Thus, x = 2
or x = 4

Exercise 1.25
|x|
What is the range of the function f (x) = x
for all x 6= 0?

Solution.
|x| x |x|| x
If x > 0 then x
= x
= 1. If x < 0 then x
= x
= 1. Thus, the range is
{1, 1}

Exercise 1.26
Solve 3 |x 2| 7. Write your answer in interval notation.

Solution.
Solving the inequality |x 2| 7 we find 5 x 9. Solving the inequality
|x 2| 3 we find x 1 or x 5. Thus, the common intervals are
[5, 1] [5, 9]

7
Exercise1.27
2
Simplify |x|x .

Solution.
x2 |x|
The answer is |x|
= |x|
=1

Exercise 1.28  
Solve the inequality  x+1
x2
 < 3. Write your answer in interval notation.

Solution.
We have 3 < x+1x2
< 3. The inequality x+1
x2
< 3 implies 2x+7
x2
< 0. Solving
7
this inequality we find x < 2 or x > 2 . Likewise, The inequality x+1
x2
> 3
4x5 5
implies x2 > 0. Solving this inequality we find x < 4 or x > 2. Hence, the
common interval is (, 54 ) ( 27 , )

Exercise 1.29
Suppose x and y are real numbers such that |x y| < |x|. Show that xy > 0.

Solution.
Since |x y| < |x| we have |x| < x y < |x|. Multiplying through by 1
and adding x we obtain x |x| < y < x + |x|. If x = 0 then 0 < y < 0 which
is impossible. Therefore either x > 0 or x < 0. If x > 0 then 0 < y < 2x.
Hence xy > 0. If x < 0 then 2x < y < 0 and so xy > 0

8
Solutions to Section 2
Exercise 2.1
Prove that A is bounded if and only if there is a positive constant C such
that |x| C for all x A.
Solution.
Suppose that A is bounded. Then there exist real numbers m and M such
that m x M for all x A. Let C = {|m|, |M | + 1} > 0. Then C
|m| m x M |M | + 1 C. That is, |x| C for all x A.
Conversely, suppose that |x| C for all x A and for some C > 0. Then
, by Exercise 1.4, we have C x C for all x A. Let m = C and
M =C
Exercise 2.2
Let A = [0, 1].
(a) Find an upper bound of A. How many upper bounds are there?
(b) Find a lower bound of A. How many lower bounds are there?
Solution.
(a) Any number greater than 1 is an upper bound.
(b) Any number less than 0 is a lower bound
Exercise 2.3
Consider the set A = { n1 : n N}.
(a) Show that A is bounded from above. Find the supremum. Is this supre-
mum a maximum of A?
(b) Show that A is bounded from below. Find the infimum. Is this infimum
a minimum of A?
Solution.
(a) The supremum is 1 which is also the maximum of A.
(b) The infimum is 0 which is not a minimum of A
Exercise 2.4
Consider the set A = {1 n1 : n N}.
(a) Show that 1 is an upper bound of A.
(b) Suppose L < 1 is another upper bound of A. Let n be a positive integer
1
such that n > 1L . Such a number n exist by the Archimedian property
which we will discuss below. Show that this leads to a contradiction. Thus,
L 1. This shows that 1 is the least upper bound of A and hence sup A = 1.

9
Solution.
(a) Since n1 > 0 we have 1 n1 < 1 for all n N. Thus, 1 is an upper bound
of A.
1
(b) Since n > 1L , we find L < 1 n1 . But this contradicts the fact that L is
an upper bound of A. Hence, we must have 1 < L. This shows, that 1 is the
smallest upper bound and so is the supremum of A

Exercise 2.5
Let a, b R with a > 0.
(a) Suppose that na b for all n N. Show that the set A = {na : n N}
has a supremum. Call it c.
(b) Show that na c a for all n N. That is, c a is an upper bound of
A. Hint: n + 1 N for all n N.
(c) Conclude from (b) that there must be a positive integer n such that
na > b.

Solution.
(a) Since na b for all n N, the set A is bounded from above. By the
completeness axiom of R, A has a supremum, denote it by c = sup{A}.
(b) Let n N. Then n + 1 N so that (n + 1)a c or na c a.
(c) From (b), we have that c a is an upper bound of A. By the definition
of supremum, we must have c c a which is impossible. This shows that
A cannot have an upper bound. That is, there must be a positive integer n
such that na > b

Exercise 2.6
Let a and b be two real numbers such that a < b.
(a) Let [a] denote the greatest integer less than or equal to a. Show that
[a] 1 < a < [a] + 1.
1
(b) Let n be a positive integer such that n > ba . Show that na + 1 < nb.
(c) Let m = [na] + 1. Show that na < m < nb. Thus, a < m n
< b. We see
that between any two distinct real numbers there is a rational number.

Solution.
(a) Since |[a] a| < 1, we have 1 < [a] a < 1 which is equivalent to
[a] 1 < a < [a] + 1.
1
(b) Since n > ba and b a > 0 we can have n(b a) > 1 or na + 1 < nb.
(c) We have m 1 = [na] na < [na] + 1 = m < na + 1 < nb. Thus,
na < m < nb. Dividing through by n > 0 we obtain a < m n
<b

10
Exercise 2.7 n
Consider the set A = { (1)
n
: n N}.
(a) Show that A is bounded from above. Find the supremum. Is this supre-
mum a maximum of A?
(b) Show that A is bounded from below. Find the infimum. Is this infimum
a minimum of A?
Solution.
(a) The supremum is 12 which is also a maximum.
(b) The infimum is 1 which is also a minimum
Exercise 2.8
Consider the set A = {x R : 1 < x < 2}.
(a) Show that A is bounded from above. Find the supremum. Is this supre-
mum a maximum of A?
(b) Show that A is bounded from below. Find the infimum. Is this infimum
a minimum of A?
Solution.
(a) 2 is a supremum that is not a maximum.
(b) 1 is an infimum that is not a minimum
Exercise 2.9
Consider the set A = {x R : x2 > 4}.
(a) Show x A and x < 2 leads to a contradiction. Hence, we must have
that x 2 for all x A. That is, 2 is a lower bound of A.
(b) Let L be a lower bound of A such that L > 2. Let y = L+22
. Show that
2 < y < L.
(c) Use (a) to show that y A and L y. Show that this leads to a
contradiction. Hence, we must have L 2 which means that 2 is the infimum
of A.
Solution.
(a) If x A and x < 2 then x2 < 4 which contradicts the fact that x A.
Thus, for all x A we have x 2. This shows that 2 is a lower bound of A.
(b) Since L > 2 we have L + 2 > 4 and this implies y = L+2 2
> 2. Also,
L+2 L+L
y = 2 < 2 = L.
(c) Since y > 2 we have y 2 > 4 so that y A. But L is a lower bound of A
so we must have L y. But this contradicts y < L from (b). It follows that
2 is the least lower bound of A

11
Exercise 2.10
Show that for any real number x there is a positive integer n such that n > x.

Solution.
Let a = 1 and b = x in the Archimedean property

Exercise 2.11
Let a and b be any two real numbers such that a < b.
(a) Let w be a fixed positive irrational number. Show that there is a rational
number r such that a < wr < b.
(e) Show that wr is irrational. Hence, between any two distinct real numbers
there is an irrational number.

Solution.
(a) Since a < b, we have wa < wb . By Exercise ??, there is a rational number
r such that wa < r < wb or a < rw < b.
(e) If rw = s with s rational then w = rs which is a rational, a contradiction.
Hence, rw is irrational

Exercise 2.12
Suppose that = sup A < . Let  > 0 be given. Prove that there is an
x A such that  < x.

Solution.
Suppose the contrary. That is,  x for all x A. In this case,  is
an upper bound of A. Thus, we must have  which is impossible

Exercise 2.13
Suppose that = inf A < . Let  > 0 be given. Prove that there is an
x A such that +  > x.

Solution.
Suppose the contrary. That is, +  x for all x A. In this case, +  is
a lower bound of A. Thus, we must have +  which is impossible

Exercise 2.14
For each of the following sets S find sup{S} and inf{S} if they exist.
(a) S = {x R : x2 < 5}.
(b) S = {x R : x2 > 7}.
(c) S = { n1 : n N}.

12
Solution.
(a) sup{S} = 5 and inf{S} = 5.
(b) sup{S} = and inf{S} = .
(c) sup{S} = 0 and inf{S} = 1

13
Solutions to Section 3
Exercise 3.1
Find a simple expression for the general term of each sequence.
(a) 1, 12 , 31 , 14 ,
(b) 2, 23 , 43 , 54 ,
(c) 1, 13 , 51 , 17 , 19 ,
(d) 1, 1, 1, 1, 1, 1,

Solution.
(a) an = (1)n1 n1 .
(b) an = n+1
n
.
1
(c) an = 2n1
(d) an = (1)n

Exercise 3.2 
Show that the sequence n1 n=1 converges to 0.

Solution.
Let  > 0 be given. We want to find a positive integer N such that if n N
then  n1 0 < . But this last inequality implies that n > 1 . Let N be a
positive integer greater than 1 . If n N > 1 then  n1 0 = n1 < . This
shows that
1
lim = 0
n n

Exercise 3.3
Show that the sequence 1 + Cn n=1 converges to 1, where C 6= 0 is a con-


stant.

Solution.  
Let  > 0. We want to find a positive integer N such that 1 + Cn 1 < 
for all n N . But 1 + Cn 1 <  implies n > |C|
 

. By choosing N to be a
|C|
positive integer greater than  the result follows

Exercise 3.4
Is there a number L with the property that |(1)n L| < 1 for all n N1 ,
where N1 is some positive integer? Hint: Consider the inequality with an
even integer greater than N1 and an odd integer greater than N1 .

14
Solution.
If ne N1 is an even integer than |(1)ne L| = |1 L| < 1. If no N1 is an
odd integer than |(1)no L| = | 1 L| < 1. This shows that L is within
one unit of both 1 and 1 which is impossible. Thus, L does not exist
Exercise 3.5
Use the previous exercise to show that the sequence {(1)n }
n=1 is divergent.

Solution.
Assume the contrary. That is, suppose there is an L such that limn (1)n =
L. Let  = 1. Then, there is a positive integer N1 such that n N1 implies
|(1)n L| < 1. But by the previous exercise, this is impossible. Hence, the
given sequence is divergent
Exercise 3.6
Suppose that limn an = a and limn an = b with a < b. Show that by
choosing  = ba
2
> 0 we end up with the impossible inequality b a < b a.
A similar result holds if b < a. Thus, we must have a = b. Hint: Exercise 1.6
and Exercise 1.17.
Solution.
Let  = ba
2
> 0. Since the sequence converges to a, we can find a positive
integer N1 such that
ba
n N1 = |an a| < .
2
Similarly, since the sequence converges to b we can find a positive integer N2
such that
ba
n N2 = |an b| < .
2
Let N = max{N1 , N2 }. Then for n N we have n N1 and n N2 .
Moreover, by using Exercise 1.6 and Exercise 1.17 we have
b a =|b a| = |(b an ) + (an a)|
|b an | + |an a|
ba ba
=|an b| + |an a| < + =ba
2 2
Thus, we conclude that b a < b a which is impossible. Likewise, if b < a
we end up with a b < a b which is impossible. That is, either a < b or
b < a leads to a contradiction. Hence, a = b

15
Exercise 3.7
Show that each of the following sequences is bounded. Identify M in each
case.
(a) an = (1)n .
(b) an = n ln1(n+1) .

Solution.
n
(a) We have |a n | = |(1) | = 1 1 so that M = 1.
(b) n 1 n 1 1n 1. Also, ln (n + 1) ln 2 1
ln (n+1)
1
ln 2
.
Hence, |an | ln12 = M

Exercise 3.8
Let {an }
n=1 be a sequence such that |an | K for all n N. Show that this
sequence is bounded. Identify your M.

Solution.
Let M = |a1 | + |a2 | + + |aN 1 | + K. Then |an | M for all n 1. That
is, the sequence is bounded

Exercise 3.9
Show that a convergent sequence is bounded. Hint: use the definition of
convergence with  = 1.

Solution.
Let {an }n=1 be a convergent sequence with limit L. Let  = 1. There is a
positive integer N1 such that |an L| < 1 for all n N1 . By Exercise 1.18,
we obtain |an | |L| < 1 or |an | < 1 + |L| for all n N1 . By the previous
problem with K = 1 + |L|, the sequence is bounded

Exercise 3.10
Give an example of a bounded sequence that is divergent.

Solution.
Let an = (1)n . We know that this sequence is bounded (Exercise 3.7(a)).
We also know that this sequence is divergent (Exercise )

Exercise 3.11
Let {an }
n=1 , {bn }n=1 , {cn }n=1 be three sequences with the following condi-
tions:

16
(1) bn an cn for all n K, where K is some positive integer.
(2) limn bn = limn cn = L.
Show that limn an = L. Hint: Use the definition of convergence along
with Exercise 1.14.

Solution.
Let  > 0. By hypothesis, there exist positive integers N1 and N2 such that
|bn L| <  for all n N1 or equivalently L  < bn < L +  for all n N1
and
|cn L| <  for all n N2 or equivalently L  < cn < L +  for all n N2 .
Let N = N1 + N2 + K. Suppose n N. Then L  < bn an cn < L + .
That is L  < an < L +  for all n N. This implies that limn an = L

Exercise 3.12
An expansion of (a+b)n , where n is a positive integer is given by the Binomial
formula n
X
n
(a + b) = C(n, k)ak bnk
k=0
n!
where C(n, k) = k!(nk)! .
(a) Use the Binomial formula to establish the inequality
1 x
(1 + x) n 1 + , x0
n
1
(b) Show that if a 1 then limn a n = 1. Hint: Use Exercise 3.3.

Solution.
(a) Using the Binomial formula with a = 1 and b = nx we find

n
1 + nx = 1 + n nx +(other positive terms) 1 + x.
or  x n
1+x 1+ .
n
Taking the nth root of both sides we find
1 x
(1 + x) n 1 + .
n
17
 1
(b) If a 1 then a n 1. By letting x = a 1 0 in (a) we find
1 a1
1 an 1 + .
n
By Exercise 3.3 we know that limn 1 + a1
n
= 1. Thus, by the squeeze rule
we obtain
1
lim a n = 1, a 1
n

Exercise 3.13
Prove that the sequence {cos (n)}
n=1 is divergent.

Solution.
Note that {cos (n)} n
n=1 = {(1) }n=1 and by Exercise , this sequence is
divergent

Exercise 3.14
Let {an }
n=1 be the sequence defined by an = n for all n N. Explain why
the sequence {an }
n=1 does not converge to any limit.

Solution.
The sequence is unbounded

Exercise 3.15
(a) Show that for all n N we have
n! 1
n
.
n n
(b) Show that the sequence {an }
n=1 where an =
n!
nn
is convergent and find
its limit.

Solution.
n(n1)(n2)21
(a) We know that ni n
1 for all 0 i n 1. Thus, n!
nn
= nnn
=
n n1 n2
n
n n n n n1 .
2 1

(b) By the Squeeze rule we find that limn nn!n = 0

Exercise 3.16
Using only the definition of convergence show that

3
n 5001
lim = 1.
n 3
n 1001

18
Solution.
Let  > 0. We want to find a positive integer N such that if n N then
 
 3 n 5001 
 1<
 3 n 1001 

or  
 4001 

 3 n 1001  < .


Let n > 10013 . Then 3
n 1001 > 0 so that the previous inequality becomes
4001
< .
3
n 1001

Solving this for n we find

 3
4001
n> + 1001 .


3
Let N be a positive integer greater than 4001

+ 1001 . Then for n N we
have  
 3 n 5001 
 1<
 3 n 1001 

Exercise 3.17
Consider the sequence defined recursively by a1 = 1 and an+1 = 2 + an for
all n N. Show that an 2 for all n N.

Solution.
The proof is by induction n = 1 we have a1 = 1 2. Suppose that
on n. For
an 2. Then an+1 = 2 + an 2 + 2 = 2

Exercise 3.18
(n2 +1) cos n
Calculate limn n3
.

Solution.
We have
n2 + 1 (n2 + 1) cos n n2 + 1
.
n3 n3 n3
By the Squeeze rule we conclude that the limit is 0

19
Exercise 3.19
2(1)n+3
Calculate limn
n
.

Solution.
We have
2 2(1)n+3 2
.
n n n
By the Squeeze rule the limit is 0

Exercise 3.20
Suppose that limn an = L with L > 0. Show that there is a positive
integer N such that 2xN > x.

Solution.
Let  = L2 . Then there is a positive integer N such that if n N we have
|an L| < L2 . Thus, |aN L| < L2 or L2 < aN L. Hence, aN > L2 or
2aN > L

Exercise 3.21
Let a R and n N. Clearly, a < a + n1 .
(a) Show that there is a1 Q such that a < a1 < a + n1 . Hint: Exercise
2.6(c).
(b) Show that there is a2 Q such that a < a2 < a1 .
(c) Continuiung the above process we can find a sequence {an } n=1 such that
1
a < an < a + n for all n N. Show that this sequence converges to a.
We have proved that if a is a real number then there is a sequence of rational
numbers converging a. We say that the set Q is dense in R.

Solution.
(a) This follows from Exercise 2.6(c).
(b) Similar to (a).
(c) Applying the Squeeze rule, we obtain limn an = a

20
Solutions to Section 4
Exercise 4.1
Suppose that limn an = A and limn bn = B. Show that

lim an bn = A B.
n

Solution.
We will prove the result for addition. The difference case is similar. Let
 > 0. Since the two sequences are convergent, there exist positive integers
N1 and N2 such that

|an A| < 2
for all n N1
and

|bn B| < 2
for all n N2 .
Let N = N1 + N2 . Then for all n N we have n N1 and n N2 . Hence,
|(an + bn ) (A + B)| = |(an A) + (bn B)| |an A| + |bn B| < 2 + 2 = .
This establishes the desired result

Exercise 4.2
Suppose that limn an = A and limn bn = B.
(a) Show that |bn | M for all n N, where M is a positive constant.
(b) Show that an bn AB = (an A)bn + A(bn B).
(c) Let  > 0 be arbitrary and K = M +|A|. Show that there exists a positive

integer N1 such that |an A| < 2K for all n N1 .
(d) Let  > 0 and K be as in (c). Show that there exists a positive integer

N2 such that |bn B| < 2K for all n N2 .
(e) Show that limn an bn = AB.

Solution.
(a) Since {bn }n=1 is convergent, the sequence is bounded. Thus, there is a
positive constant M such that |bn | M for all n 1.
(b) We have (an A)bn + A(bn B) = an bn Abn + Abn AB = an bn AB.

(c) Let 1 = 2K . Since limn an = A, we can find a positive integer N1 such

that |an A| < 1 = 2K for all n N1 .

(d) Let 2 = 2K . Since limn bn = B, we can find a positive integer N2 such

21
 
that |bn B| < 2 = 2K for all n N2 .
(e) Let N = N1 + N2 . Then n N implies n N1 and n N2 . In this case,

|an bn AB| =|(an A)bn + A(bn B)| |an A||bn | + |A||bn B|

 
< M + |A|
2K 2K
 
< + =
2 2
M M |A| |A|
where we have used the fact that K
= M +|A|
< 1 and K
= M +|A|
<1

Exercise 4.3
Give an example of two divergent sequences {an }
n=1 and {bn }n=1 such that
{an bn } and {an + bn } are convergent.

Solution.
Let an = (1)n and bn = (1)n+1 . Both sequences are divergent. Moreover,
an bn = (1)2n+1 = 1 for all n 1. Hence, limn an bn = 1. Finally,
an + bn = (1)n (1)n = 0 for all n 1. Therefore, limn (an + bn ) = 0

Exercise 4.4
Let k 6= 0 be an arbitrary constant and limn an = A. Show that limn kan =
kA.

Solution.

Let  > 0 be arbitrary. There is a positive integer N such that |an A| < |k|
for all n N. Moreover, for n N we have

|kan kA| = |k(an A)| = |k||an A| < |k| = .
|k|

This establishes the desired result

Exercise 4.5
Suppose that limn an = 0 and {bn }
n=1 is bounded. Show that limn an bn =
0.

Solution.
Since {bn }
n=1 is bounded, we can find a positive constant M such that |bn |
M for all n 1. Let  > 0 be arbitrary. Since limn an = 0, there is a

22
 
positive integer N such that |an 0| = |an | < M
for all n N. Thus, for
n N, we have

|an bn 0| = |an bn | = |an ||bn | < M = .
M
This shows that limn an bn = 0
Exercise 4.6
(a) Use the previous exercise to show that limn sinn n = 0.
(b) Show that limn sinn n = 0 using the squeeze rule.
Solution.
(a) Let an = n1 and bn = sin n. Then limn an = limn n1 = 0 and
|bn | = | sin n| 1. Now the result follows from the previous exercise.
(b) Since 1 sin n 1, we obtain n1 sinn n n1 . But limn n1 =
limn n1 = 0 so that by the squeeze rule
sin n
lim =0
n n

Exercise 4.7
Suppose that limn an = A, with A 6= 0. Show that there is a positive
integer N such that |an | > |A|
2
for all n N. Hint: Use Exercise 1.18.
Solution.
Let  = |A|
2
> 0. Then there is a positive integer N such that |an A| < |A| 2
for all n N. By Exercise 1.18, we have |A| |an | < |A|
2
or |a n | > |A|
2
for
nN
Exercise 4.8
Let {an }
n=1 be a sequence with the following conditions:
(1) an 6= 0 for all n 1.
(2) limn an = A, with A 6= 0.
(a) Show that there is a positive integer N1 such that for all n N1 we have
 
1
  < 2 |an A|.
1 
 an A  |A|2

(b) Let  > 0 be arbitrary. Show that there is a positive integer N2 such that
for all n N2 we have
|A|2
|an A| < .
2
23
(c) Using (a) and (b), show that
1 1
lim = .
n an A

Solution.
(a) From the previous exercise, we can find a positive integer N1 such that
|an | > |A|
2
for n N1 . Thus, for n N1 we obtain
 
 1  = |an A| < 2 |an A|.
1 
 an A  |an ||A| |A|2

(b) Let  > 0 be arbitrary. Since limn an = A, we can find a positive

integer N2 such that for all n N2 we have

|A|2
|an A| < .
2
(c) Let  > 0 be arbitrary. Let N = N1 + N2 . Then n N implies that
n N1 and n N2 . Moreover,
2
 
  = |an A| < 2 |an A| < 2 |A|  = .
1 1 
 an A  |an ||A| |A|2 |A|2 2

This establishes the required result

Exercise 4.9
1
Let 0 < a < 1. Show that limn a n = 1. Hint: Use Exercise 3.12 (b).

Solution.
1
Since 0 < a < 1, we have a
> 1. By Exercise 3.12(b), we find
  n1
1 1
lim = lim 1 = 1.
n a n a n

By the previous exercise, we can write

1 1 1
lim a n = lim 1 = =1
n n 1 1
an

24
Exercise 4.10
Show that if limn an = A and limn bn = B with bn 6= 0 for all n 1
and B 6= 0, then
an A
lim = .
n bn B

Solution.
Using Exercise ?? and Exercise ?? we can write
an 1 1 A
lim = lim an =A =
n bn n bn B B

Exercise 4.11
Given that limn an = A and limn bn = B with an bn for all n 1.
(a) Suppose that B < A. Let  = AB2
> 0. Show that there exist positive
integers N1 and N2 such that A  < an < A +  for n N1 and B  <
bn < B +  for n N2 .
(b) Let N = N1 + N2 . Show that for n N we obtain the contradiction
bn < an . Thus, we must have A B.

Solution.
(a) Since limn an = A, there exists a positive integer N1 such that |an
A| <  for all n N1 . By Exercise 1.14, this is equivalent to A < an < A+
for n N1 . Similarly, since limn bn = B, there exists a positive integer
N2 such that |bn B| <  for all n N2 . By Exercise 1.14, this is equivalent
to B  < bn < B +  for n N2 .
(b) If n N then n N1 and n N2 . Thus,

AB A+B
bn <B +  = B + =
2 2
AB
=A = A  < an
2
This contradicts the fact that an bn for all n 1. Hence, we conclude that
AB

Exercise 4.12
an
Suppose that limn bn
= L and limn bn = 0 where bn 6= 0 for all n N.
Find limn an .

25
Solution.
We have
 
an an 
lim an = lim bn = lim lim bn = L 0 = 0
n n bn n bn n

Exercise 4.13
The Fibonacci numbers are defined recursively as follows:
a1 = a2 = 1 and an+2 = an+1 + an for all n N.
an+1
Suppose that limn an
= L. Find the value of L.
Solution.
We have
an+2 = an+1 + an
Divide through by an+1 to obtain
an+2 an
=1+ .
an+1 an+1
Take the limit as n to obtain
1
L=1+
L
or
L2 L 1 = 0.

1+ 5
Solving this quadratic equation for L > 0 we find L = 2

Exercise 4.14
Show that the sequence defined by
n n2 + 3
an = + (1)n 2
n+1 n +7
have two limits by finding limn a2n and limn a2n+1 .
Solution. We have
2n 4n2 + 3
lim a2n = lim + 2 =2
n n 2n + 1 4n + 7
and
2n + 1 (2n + 1)2 + 3
lim a2n+1 = lim =0
n n 2n + 2 (2n + 1)2 + 7

26
Exercise 4.15
Use the properties of this section to find

2n2 + 5n
lim .
n n+4
Solution.
We have

q
5
2n2 + 5n 2+ n
lim = lim
n n+4 n 1 + 4
n
q
5
limn 2 + n
= lim
n limn (1 + n4 )
q
limn (2 + n5 )
= lim 4
n 1 + limn n

= 2
Exercise 4.16
Find the limit of the sequence defined by
1
an = n 2 ln n .
Solution.
1 1
We have limn ln an = limn 2 ln n
ln n = 12 . Thus, limn an = e 2
Exercise 4.17
Consider the sequence defined by
1 1 1
an = + + + .
1 2 n

(a) Show that an n for all n N.
(b) Show that the sequence {an }
n=1 is divergent. Hint: Exercise 4.11.

Solution.
(a) By induction on n. If n = 1 we have a1 = 11 = 1. Suppose that

an n. Then an+1 n + n+1 1
= nn+1
n+1+1 n n+1
= n+1 = n + 1.
n+1 n+1
(b) Since an n for all n N and limn n = , we conclude that
limn an =

27
Exercise 4.18
Find the limit of the sequence defined by

an = ln (2n + n) ln n.

Solution.    
We have limn [ln (2n + n)ln n] = limn ln 2n+n n = limn ln 2 + 1
n
=
ln 2

Exercise 4.19
Consider the sequence defined
by an = n 3n + 1.
(a) Show that 3 < an < 3 n 2 for all n 2.
(b) Find the limit of an as n .

Solution.
(a) We have
n n
3 = 3n < n 3n + 1 < n 3n + 3n < 3 2.

(b) Since limn n 2 = 1 we conclude by the Squeeze rule that limn an =
3

Exercise 4.20
Let {an }
n=1 be a convergent sequence of nonnegative terms with limit L.
Suppose that the terms of sequence satisfy the recursive relation an an+1 =
an + 2 for all N N. Find L.

Solution.
We have limn an an+1 = limn (an + 2) limn an limn an+1 =
limn an + limn 2 L L = L + 2 L2 L 2 = 0. Solving this
equation we find L = 1 and L = 2. But an 0 for all n N so that L 0.
Thus, L = 2

Exercise 4.21
Find the limit of the sequence defined by
1 sin n
an = cos + .
n n
Solution.
We have
sin n
lim an = lim cos 1n + lim =1+0=1
n n n n

28
Exercise 4.22
a2n +1
Suppose that an+1 = an
. Show that the sequence {an }
n=1 is divergent.

Solution.
2
Suppose the contrary and let L = limn an . Then L = L L+1 or L2 = L2 + 1
which leads to the contradiction 0 = 1. Hence, the given sequence must be
divergent

29
Solutions to Section 5
Exercise 5.1
Show that the sequence { n1 }
n=1 is decreasing.

Solution.
1 1
We have n + 1 n n+1 n
an an+1 for all n 1. This shows that
the sequence is decreasing

Exercise 5.2
Show that the sequence { 1+e1n }
n=1 is increasing.

Solution.
We have n n + 1 en en+1 e(n+1) en 1 + e(n+1) 1 + en
1 1
1+en
1+e(n+1) an an+1 for all n 1. This shows that the sequence
is increasing

Exercise 5.3
Show that the sequence { n1 }
n=1 is bounded from below. What is a lower
bound? Are there more than one lower bound?

Solution.
Since n1 0 for all n 1, the given sequence is bounded from below with a
lower bound 0. Any negative number is a lower bound

Exercise 5.4
Show that the sequence { 1+e1n }
n=1 is bounded from above. What is an upper
bound? Are there more than one upper bound?

Solution.
We have 1 + en 1 1+e1n 1. Thus, the given sequence is bounded
from above with an upper bound equals to 1. Any number larger than 1 is
an upper bound

Exercise 5.5
Let {an }
n=1 be an increasing sequence that is bounded from above.
(a) Show that there is a finite number M such that M = sup{an : n 1}.
(b) Let  > 0 be arbitrary. Show that M  cannot be an upper bound of
the sequence.

30
(c) Show that there is a positive integer N such that M  < aN .
(d) Show that M  < an for all n N.
(e) Show that M  < an < M +  for all n N.
(f) Show that limn an = M. That is, the given sequence is convergent.

Solution.
(a) Since the sequence {an }n=1 is bounded from above, by the Completeness
Axiom there is a finite number M such that M = sup{an : n 1}. We will
show that {an }n=1 converges to M.
(b) Let  > 0 be arbitrary. Consider the number M . This number is
either an upper bound of the sequence or not. If it is an upper bound then
we must have M M  which is an impossible inequality.
(c) If such an N does not exist, then we will have an M  for all n 1.
This means that M  is an upper bound which is impossible by (b). Thus,
there is a positive integer N such that

M  < aN M.

(d) Since the sequence is increasing, for all n N we can write aN an .

Hence, for all n N we have M  < an .
(e) For all n N we can write

M  < an M < M + 

or equivalently |an M | <  for all n N.

(f) This follows from the definition of convergence and (e)

Exercise 5.6
Consider the sequence {an } 3
n=1 defined recursively by a1 = 2 and an+1 =
1
a + 1 for n 2.
2 n
1
(a) Show by induction on n 2, that an+1 = an + 2n+1 .
(b) Show that this sequence is increasing.
(c) Show that {an }n=1 is bounded from above. What is an upper bound?
(d) Show that {an }n=1 is convergent. What is its limit? Hint: In finding the
limit, use the arithmetic operations of sequences.

Solution.
(a) We have a2 = 21 a1 + 1 = 34 + 1 = 74 = 32 + 41 = a1 + 212 . Suppose

1that
1 1 1 1 1
an = an1 + 2n . Then an+1 = 2 an +1 = 2 an1 + 2n+1 +1 = 2 an1 + 1 + 2n+1 =

31
 1
an + 2n+1 .
1
(b) Since an+1 an = 2n+1 > 0, the given sequence is increasing.
(c) By (b), we have an+1 = 21 an + 1 > an . Solving this equality for an we find
an < 2 so that the sequence {an } n=1 is bounded from above with an upper
bound equals to 2.
(d) By the previous exercise, the sequence is convergent say to A. Using the
arithmetic operations of sequences, we can write
 
1 1
lim an+1 = lim an + 1 = lim an + lim 1.
n n 2 2 n n

Thus,
1
A = A + 1.
2
Solving this equation for A we find A = 2

Exercise 5.7
Let {an }
n=1 be a decreasing sequence such that M an for all n 1. Show
that {an }
n=1 is convergent. Hint: Let bn = an and use Exercise 5.5 and
Exercise 4.4.

Solution.
Since {an }
n=1 is decreasing, the sequence {bn }n=1 is increasing. Since the
sequence {an }n=1 is bounded from below, the sequence {bn }

n=1 is bounded

from above. By Exercise 5.5, the sequence {bn }n=1 is convergent. By Exercise
4.4, the sequence {an }n=1 is also convergent

Exercise 5.8
Show that a monotone sequence is convergent if and only if it is bounded.

Solution.
Suppose first that a sequence {an } n=1 is a monotone convergent sequence.
By Exercise 3.9, the sequence {an }n=1 is bounded.

Conversely, suppose that {an }n=1 is a bounded sequence. Then there is a
positive constant M such that |an | M for all n 1. By Exercise 1.14,
we have M an M for all n 1. This shows that the sequence is
bounded from below as well from above. If the sequence is either increasing
or decreasing, then it is convergent by Exercise 5.5 and Exercise 5.7

32
Exercise 5.9
Let an be defined by a1 = 2 and an+1 = 2 + an for n N.
(a) Show that an 2 for all n N. That is, {an }
n=1 is bounded from above.
(b) Show that an+1 an for all n N. That is, {an } n=1 is increasing.
(c) Conclude that {an }
n=1 is convergent. Find its limit.

Solution.
n. For n =1, we have a1 = 2 2. Suppose that
(a) By induction on
an 2. Then an+1 = 2 + an 2 + 2 = 2. Thus, an N.
p2 forall n
(b) By induction on n. For n = 1 we have a2 a1 = 2 + 2 2 0.
Suppose that an an1 0. Then
p an an1
an+1 an = 2 + an 2 + an1 = 0.
2 + an + 2 + an1
(c) Since the sequence is increasing and
bounded from above, it is convergent,
say with limit a. Thus, we have a = 2 + a. Solving this equation we obtain
a = 1 or a = 2. Since a1 = 2 and the sequence is increasing we conclude
that a = 2

ExerciseP 5.10
Let an = nk=1 k12 .
(a) Show that an < 2 for all n N. Hint: Recall that nk=1 1 1
P
(n+1)n
= 1 n+1 .
(b) Show that {an }n=1 is increasing.
(c) Conclude that {an }n=1 is convergent.

Solution.
(a) We have an = nk=1 k12 = 1 + nk=2 k12 1 + 1 n1 < 2.
P P
1
(b) Since an+1 = an + (n+1)2 > an , the given sequence is increasing.

(c) This follows from the fact that an increasing sequence that is bounded
from above is convergent

Exercise 5.11
Consider the sequence {an }
n=1 defined recursively as follows

a1 = 2 and 7an+1 = 2a2n + 3 for all n N.

(a) show that 21 < an < 3 for all n N.
(b) Show that an+1 an for all n N.
(c) Deduce that {an }
n=1 is convergent and find its limit.

33
Solution.
(a) We prove this by induction on n. If n = 1 then 21 < 2 = a1 < 3. Suppose
that 12 < an < 3. Then 72 < 7an+1 < 21 12 < an+1 < 3.
(b) We have an+1 an = 71 (2a2n 7an + 3) = 17 (2an 1)(an 3) < 0. Thus,
an+1 an for all n N.
(c) We deduce that the given sequence is convergent, say with limit L. Thus,
7L = 2L2 + 3. Solving this equation for L we find L = 21 and L = 3. Since
the sequence is decreasing and 12 < an < 3 we must have L = 12

Exercise 5.12
a1 +a2 ++an
Let {an }
n=1 be an increasing sequence. Define bn = n
. Show that
the sequence {bn }
n=1 is increasing.

Solution.
note that s1 s2 sn sn+1 , hence s1 + s2 + + sn nsn+1 . Adding
n(s1 + + sn ) to both sides, we obtain
n(s1 + + sn ) + (s1 + + sn ) n(s1 + + sn ) + nsn+1 ,
or, in other words,
(n + 1)(s1 + + sn ) n(s1 + + sn + sn+1 ),
Dividing both sides by n(n + 1), we obtain
s1 + + sn s1 + + sn + sn+1
,
n n+1
or, in other words, bn bn+1 for all n. This proves our claim

Exercise 5.13
Give an example of a monotone sequence that is divergent.

Solution.
One example is the sequence defined by an = n

Exercise 5.14
an
Consider the sequence defined recursively by a1 = 1 and an+1 = 3 + 2
for
all n N.
(a) Show that an 6 for all n N.
(b) Show that {an }
n=1 is increasing.
(c) Conclude that the sequence is convergent. Find its limit.

34
Solution.
(a) If n = 1 we have a1 = 1 < 6. Suppose that an 6. Then an+1 = 3 + a2n
3 + 62 = 6.
(b) We have an+1 = 3 + a2n = 6+a 2
n
an +a
2
n
= an
(c) Since the sequence is increasing and bounded from above it is convergent
with limit say equals to L. Thus, L = 3 + L2 and solving for L we find L = 6

Exercise 5.15
Give an example of two monotone sequences whose sum is not monotone.

Solution.
Let {an }
n=1 = {1, 1, 2, 2, 3, 3, } and {bn }n=1 = {1, 2, 2, 3, 3, }.
The first sequence is increasing and the second sequence is decreasing. How-
ever the sum is the sequence {0, 1, 0, 1, 0, 1, } which is not monotone

35
Solutions to Section 6
Exercise 6.1
Let {ank }
k=1 be a subsequence of a sequence {an }n=1 . Use induction on k to
show that nk k for all k N.

Solution. For k = 1 we have n1 N so that n1 1. Suppose that nk k.

Then nk+1 > nk k. Thus, nk+1 k + 1

Exercise 6.2
Let {an }
n=1 be a sequence of real numbers that converges to a number L.
Let {ank }
k=1 be any subsequence of {an }n=1 .
(a) Let  > 0 be given. Show that there is a positive integer N 0 such that if
n N 0 then |an L| < .
(b) Let N be the first positive integer such that nN N 0 . Show that if k N
then |ank L| < . That is, the subsequence {ank }k=1 converges to L. Hence,
every subsequence of a convergent sequence is convergent to the same limit
of the original sequence.

Solution.
(a) This follows from the fact that the sequence {an }
n=1 converges to L.
(b) Let N be such that nN N . Then if k N we have nk > nN N 0 .
0

Hence, |ank L| < . This shows that the subsequence {ank }

k=1 converges to
L

Exercise 6.3
Let {an }n=1 be a sequence of real numbers. Let S = {n N : an > am for all
m > n}.
(a) Suppose that S is infinite. Then there is a sequence n1 < n2 < n3 <
such nk S. Show that ank+1 < ank . Thus, the subsequence {ank } k=1 is
decreasing.
(b) Suppose that S is finite. Let n1 be the first positive integer such that
n1 6 S. Show that the subsequence {ank }k=1 is increasing.

Solution.
(a) By the definition of S we have ank+1 < ank since nk+1 > nk .
(b) Let n1 be the first positive integer such that n1 6 S. This means that
there is a positive integer n2 > n1 such that an1 < an2 . But n2 6 S so that
there is a positive integer n3 > n2 such that an2 < an3 . Continuing this
process we find an increasing subsequence {ank }k=1

36
Exercise 6.4 (Bolzano-Weierstrass)
Every bounded sequence has a convergent subsequence. Hint: Exercise 5.8
Solution.
By the previous exercise, the sequence has a bounded monotone subsequence.
By Exercise 5.8 this subsequence is convergent
Exercise 6.5
Show that the sequence {esin n }
n=1 has a convergent subsequence.

Solution.
Since | sin n| 1 we must have 1e esin n e for all n N. Thus, the given
sequence is bounded so that by the Bolzano-Weierstrass theorem it has a
convergent subsequence
Exercise 6.6
Prove that the sequence {an } n
n=1 where an = cos 2 is divergent.

Solution.
Consider the two subsequences {a2n+1 }
n=1 that converges to 0 and {a4 }n=1
that converges to 1. Thus, the original sequence must be divergent
Exercise 6.7
Prove that the sequence {an }
n=1 where

(n2 + 20n + 35) sin n3

an =
n2 + n + 1
has a convergent subsequence. Hint: Show that {an }
n=1 is bounded.

Solution.
We have
 2
 (n + 20n + 35) sin n3 

|an | =  
n2 + n + 1 
(n2 + 20n + 35)| sin n3 |
=
n2 + n + 1
2
n + 20n + 35
2
n +n+1
2
n + 20n + 35 20 35
2
=1+ + 2 56
n n n
Hence, by the Bolzano Weierstraa theorem the given sequence has a conver-
gent subsequence

37
Exercise 6.8
Show that the sequence defined by an = 2 cos n sin n has a convergent
subsequence.

Solution.
We have 2 2 cos n 2 and 1 sin n 1. Adding we obtain 3
an 3 so that the given sequence is bounded. By the Bolzano-Weierstrass
theorem, the given sequence has a convergent subsequence

Exercise 6.9
True or false: There is a sequence that converges to 6 but contains a subse-
quence converging to 0. Justify your answer.

Solution.
By Exercise 6.2, this cannot happen

Exercise 6.10
Give an example of an unbounded sequence with a bounded subsequence.

Solution.
Let 
0 if n is odd
an =
n if n is even
Then {an }
n=1 is unbounded. However, the subsequence {a2n+1 }n=1 = {0, 0, 0, }
is bounded

Exercise 6.11
Show that the sequence {(1)n }
n=1 is divergent by using subsequences.

Solution.
This sequence has two subsequences {a2n }
n=1 that converges to 1 and {a2n+1 }n=1
that converges to 1. Thus, the original sequence cannot be convergent

38
Solutions to Section 7
Exercise 7.1
Consider the sequence whose nth term is given by an = n1 . Let  > 0 be
arbitrary and choose N > 2 . Show that for m, n N we have |am an | < .
That is, the above sequence is a Cauchy sequence. Hint: Exercise 1.17.

Solution.
Let n, m N. Then by Exercise 1.17 we have
 
1
 1  1 + 1 <  +  == 

n m n m 2 2

Exercise 7.2
Show that any Cauchy sequence is bounded. Hint: Let  = 1 and use Exercise
1.18.

Solution.
Let {an }n=1 be a Cauchy sequence. Let  = 1. There is a positive integer N
such that whenever n, m N we have |an am | < 1. In particular, letting
m = N we can write |an aN | < 1 for all n N. By Exercise1.18 we can
write |an | |aN | < 1 for all n N or |an | < 1 + |aN | for all n N. Let
M = max{|a1 |, |a2 |, , |aN 1 |, |aN | + 1}. Then |an | M for all n 1. That
is, {an }
n=1 is bounded

Exercise 7.3
Show that if limn an = A then {an }
n=1 is a Cauchy sequence. Thus, every
convergent sequence is a Cauchy sequence.

Solution.
Let  > 0 be arbitrary. Then there is a positive integer N such that |an A| <

2
for all n N. Thus, for all n, m N we have

|an am | =|(an A) + (A am )| |an A| + |am A|

 
< + = .
2 2
This shows that {an }
n=1 is a Cauchy sequence

39
Exercise 7.4
(a) Using Exercise 7.2, show that for each n 1, the sequence {an , an+1 , }
is bounded.
(b) Show that for each n 1 the infimum of {an , an+1 , } exists. Call it
bn .

Solution.
(a) By Exercise 7.2, there is a positive constant M such that |an | M for
all n 1. In particular, |an | M, |an+1 | M, |an+2 | M, . That is,
{an , an+1 , } is bounded.
(b) Since {an , an+1 , } is bounded, it is bounded from below. By the Com-
pleteness Axiom, bn = inf{am : m n} exists

Exercise 7.5
(a) Show that the sequence {bn }
n=1 is bounded from above.
(b) Show that the sequence {bn }n=1 is increasing. Hint: Show that bn is a
lower bound of the sequence {an+1 , an+2 , }.

Solution.
(a) By the definition of infimum, we have bn an for all n 1. Thus,
bn an |an | M. This, shows that {bn }n=1 is bounded from above.
(b) We know that bn am for all m n. In particular bn am for all
m n + 1. This shows that bn is a lower bound of {an+1 , an+2 , }. But
bn+1 is the greatest lower bound of {an+1 , an+2 , }. Hence, bn bn+1 and
the sequence {bn }
n=1 is increasing

Exercise 7.6
Show that the sequence {bn }
n=1 is convergent. Call the limit B.

Solution.
This follows from Exercise 5.5

Exercise 7.7
(a) Let  > 0 be arbitrary. Using the definition of Cauchy sequences and
Exercise ??, show that there is a positive integer N such that aN 2 < an <
aN + 2 for all n N.
(b) Using (a), show that aN 2 is a lower bound of the sequence {aN , aN +1 , }.
Thus, aN 2 bn for all n N.
(c) Again, using (a) show that bn aN + 2 for all n N. Thus, combining

40
(b) and (c), we obtain aN 2 bn < aN + 2 .
(d) Using Exercise 4.11, show that aN 2 B aN + 2 .
(e) Using (a), (d), and Exercise 1.17, show that limn an = B. Thus, every
Cauchy sequence is convergent.

Solution.
(a) Let  > 0 be arbitrary. Since {an }
n=1 is Cauchy, there is a positive integer

N such that |an am | < 2 for all n, m N. In particular, for all n N we
have |an aN | < 2 which, by Exercise 1.14, we have aN 2 < an < aN + 2
for all n N.
(b) From part (a), we have aN 2 < an for all n N. This shows, that
aN 2 is a lower bound of the sequence {an , an+1 , }
n=N for n N. But
bn is the greatest lower bound of the sequence {an , an+1 , }
n=N for n N.
We conclude that aN 2 bn for all n N.
(c) For n N we have bn an < aN + 2 .
(d) Taking the limit as n we obtain aN 2 B aN + 2 .
(e) Using (a), (d) and Exercise 1.17, we have that for n N

|an B| =|(an aN ) + (aN B)|| |an aN | + |aN B|

 
< + =
2 2
Exercise 7.8
(a) Show that if {an } 2
n=1 is Cauchy then {an }n=1 is also Cauchy.
(b) Give an example of Cauchy sequence {a2n }
n=1 such that {an }n=1 is not
Cauchy.

Solution.
(a) Since {an }n=1 is Cauchy then it is convergent. Since the product of two
convergent sequences is convergentm the sequence {a2n } n=1 is convergent and
therefore is Cauchy.
(a) Let an = (1)n for all n N. The sequence {an } n=1 is not Cauchy since
it is divergent. However, the sequence {a2n }n=1 = {1, 1, } converges to 1
so it is Cauchy

Exercise 7.9
Let {an }
n=1 be a Cauchy sequence such that an is an integer for all n N.
Show that there is a positive integer N such that an = C for all n N,
where C is a constant.

41
Solution.
Let  = 21 . Since {an }
n=1 is Cauchy, there is a positive integer N such that if
m, n N we have |am an | < 21 . But am an is an integer so we must have
an = aN for all n N

Exercise 7.10
Let {an }
n=1 be a sequence that satisfies

|an+2 an+1 | < c2 |an+1 an | for all n N

where 0 < c < 1.
(a) Show that |an+1 an | < cn |a2 a1 | for all n 2.
(b) Show that {an }
n=1 is a Cauchy sequence.

Solution.
(a) The proof is by induction on n. For n = 2 we have |a3 a2 | < c2 |a2 a1 |.
Suppose that |an+2 an+1 | < cn+1 |a2 a1 |. Then |an+3 an+2 | < c2 |an+2
an+1 | < cn+2 |a2 a1 |.
(b) Let  > 0 be given. Since limn cn = 0 we can find a positive integer
N such that if n N then |c|n < (1 c). Thus, for n > m N we have

|an am | |am+1 am | + |am+2 am+1 | + + |an an1 |

<cm |a2 a1 | + cm+1 |a2 a1 | + + cn1 |a2 a1 |
<cm (1 + c + c2 + )|a2 a1 |
cm
= |a2 a1 | < 
1c
It follows that {an }
n=1 is a Cauchy sequence

Exercise 7.11
What does it mean for a sequence {an }
n=1 to not be Cauchy?

Solution.
A sequence {an } n=1 is not a Cauchy sequence if there is a real number  > 0
such that for all positive integer N there exist n, m N such that n, m N
and |an am | 

Exercise 7.12
Let {an }
n=1 and {bn }n=1 be two Cauchy sequences. Define cn = |an bn |.

Show that {cn }n=1 is a Cauchy sequence.

42
Solution.
Let  > 0 be given. There exist positive integers N1 and N2 such that if
n, m N1 and n, m N2 we have |an am | < 2 and |bn bm | < 2 . Let
N = N1 + N2 . If n, m N then |cn cm | = ||an bn | |am bm ||
|(an bn ) + (am bm )| |an am | + |bn bm | < . Hence, {cn }
n=1 is a
Cauchy sequence

Exercise 7.13
Suppose {an }
n=1 is a Cauchy sequence. Suppose an 0 for infinitely many
n and an 0 for infinitely many n. Prove that limn an = 0.

Solution.
Let  > 0 be given. Since {an }
n=1 is Cauchy, there is a positive integer N
such that if n, m N then |an am | < . Let n N. Then the element an
is positive, negative or zero.
Case 1: Suppose an 0. Since am 0 for infinitely many m, there is
m N such that am 0 (else, there would be less than N nonpositive
terms in the sequence, which contradicts the assumption). Since n, m N,
an an am = |an am | < .
Case 2: Suppose an < 0. Since am geq0 for infinitely many m, there is m N
such that am 0 (else, there would be less than N non-negative terms in
the sequence, which contradicts the assumption). But then, since n, m
N, an < an an + am |an am | < . So, in any case, for any given  we
can find N N such that if n N then |an | < , that is, limn an = 0

Exercise 7.14
Explain why the sequence defined by an = (1)n is not a Cauchy sequence.

Solution.
We know that every Cauchy sequence is convergent. We also know that the
given sequence is divergent. Thus, it can not be Cauchy

43
Solutions to Section 8
Exercise 8.1
x2 1
Show that limx1 x1
= 2.
Solution.
Let  > 0 be arbitrary. We note first that
 2   
x 1   (x 1)(x + 1) 
 x 1 2 =  2 = |x + 1 2| = |x 1|.
  
x1
Thus, choose . If 0 < |x 1| < (which is ) we obtain
 2 
x 1 
 x 1 2 = |x 1| < 
 

which is the required result

Exercise 8.2
Let f (x) = |x|
x
. Suppose that limx0 f (x) = L.
(a)
 Show
 that there is a positive number such that if 0 < |x| < then
 |x|  1
 x L < 4 .
(b) Let x1 = 4 and x2 = 4 . Compute the value of |f (x1 ) f (x2 )|.
(c) Use (a) to show that |f (x1 ) f (x2 )| < 21 .
(d) Conclude that L does not exist. That is, limx0 |x|
x
does not exist.
Solution.  
1  |x|
(a) Since limx0 f (x) = L, for  = we can find a > 0 such that  x L <

4
1
4
whenever 0 < |x 0| < .
(b) We have |f (x1 ) f (x2 )| = |1 (1)| = 2.
(c) Since both x1 and x2 satisfy 0 < |x1 | < and 0 < |x2 | < , we have
|f (x1 ) L| < 14 and |f (x2 ) L| < 14 . Thus,
|f (x1 ) f (x2 )| =|(f (x1 ) L) (f (x2 ) L)| (1)
|f (x1 ) L| + |f (x2 ) L|
1 1 1
< + =
4 4 2
But by (b), we have that |f (x1 ) f (x2 )| = 2. We conclude that 2 < 21 which
in impossible.
(d) The contradiction obtained in (c) shows that limx0 |x| x
does not exist

44
Exercise 8.3
Let f (x) = sin x1 . Suppose that limx0 f (x) = L.

(a)
 Show  there is a positive number such that if 0 < |x| < then

that
sin 1 L < 1 .
x 4
2 1
(b) Let n be a positive integer such that x1 = (2n+1) < and x2 = (2n+1) <
. Compute the value of |f (x1 ) f (x2 )|.
(c) Use (a) to show that |f (x1 ) f (x2 )| < 12 .
(d) Conclude that L does not exist. That is, limx0 sin x1 does not exist.

Solution.
1
(a)
 Since lim
 x0 f (x) = L, for  = 4 we can find a > 0 such that
sin 1 L < 1 whenever 0 < |x 0| < .


x 4
(b) We have |f (x1 ) f (x2 )| = | 1 0| = 1.
(c) Since both x1 and x2 satisfy 0 < |x1 | < and 0 < |x2 | < , we have
|f (x1 ) L| < 14 and |f (x2 ) L| < 14 . Thus,

|f (x1 ) f (x2 )| =|(f (x1 ) L) (f (x2 ) L)| (2)

|f (x1 ) L| + |f (x2 ) L|
1 1 1
< + =
4 4 2
But by (b), we have that |f (x1 ) f (x2 )| = 1. We conclude that 1 < 21 which
in impossible.
(d) The contradiction obtained in (c) shows that limx0 sin x1 does not

exist

Exercise 8.4
Suppose that limxa f (x) exists. Also, suppose that limxa f (x) = L1 and
limxa f (x) = L2 . So either L1 = L2 or L1 6= L2 .
(a) Suppose that L1 6= L2 . Show that there exist positive constants 1 and 2
such that if 0 < |x a| < 1 then |f (x) L1 | < |L1 L
2
2|
and if 0 < |x a| < 2
|L1 L2 |
then |f (x) L2 | < 2
.
(b) Let = min{1 , 2 } so that < 1 and < 2 . Show that if 0 < |xa| <
then |L1 L2 | < |L1 L2 | which is impossible.
(c) Conclude that L1 = L2 . That is, whenever a function has a limit, that
limit is unique.

Solution.
|L1 L2 |
(a) Since L1 6= L2 , we have  = 2
> 0. Since limxa f (x) = L1 , there

45
is 1 > 0 such that if 0 < |x a| < 1 then |f (x) L1 | < |L1 L
2
2|
. Similarly,
since limxa f (x) = L2 there is 2 > 0 such that if 0 < |x a| < 2 then
|f (x) L2 | < |L1 L
2
2|
.
(b) Suppose 0 < |x a| < . Then

|L1 L2 | =|f (x) L2 + L1 f (x)|

|f (x) L1 | + |f (x) L2 |
|L1 L2 | |L1 L2 |
< + = |L1 L2 |
2 2
(c) The contradiction obtained in (b) implies L1 = L2

Exercise 8.5
Using the  definition of limit show that

lim (2x2 + x + 1) = 2.
x1

Solution.
Let  > 0 be given. Note first that |2x2 + x + 1 2| = |(2x 1)(x + 1)| =
|2x 1||x + 1| (2|x| + 1)|x + 1|. If 0 < |x + 1| < 1 then |x| 1 < 1 and
this implies |x| < 2. So choose = min{1, 5 }. Clearly, 1 and 5 . So if
0 < |x+1| < we have |2x2 +x+12| (2|x|+1)|x+1| < 5|x+1| 5 5 = 

Exercise 8.6
x
Prove directly from the definition that limx1 x+3
= 14 .

Solution.
First note that  
 x 1  3 |x 1|
 x + 3 4  = 4 |x + 3| .
 

If |x 1| < 1 then 0 < x < 2. Thus, |x + 3| = x + 3 > 3. Let  > 0 be given.

x |x1|
Choose = min{1, 4}. Then if 0 < |x 1| < we have  x+3 14  = 34 |x+3| <
3 |x1| |x1| 4
4 3
= 4
< 4
=

Exercise 8.7
In this exercise we discuss the concept of sided limits.
(a) We say that L is the left side limit of f as x approaches a from the left
if and only if

46
  > 0, > 0 such that 0 < a x < |f (x) L| < 

and we write limxa f (x) = L. Show that limx0 |x|

x
= 1.
(b) We say that L is the right side limit of f as x approaches a from the
right if and only if

 > 0, > 0 such that 0 < x a < |f (x) L| < 

|x|
and we write limxa+ f (x) = L. Show that limx0+ x
= 1.

Exercise 8.8
Prove that L = limxa f (x) if and only if limxa f (x) = limxa+ f (x) = L.

Solution. Suppose first that limxa f (x) = L. Let  > 0 be given. Then
there is > 0 such that 0 < |x a| < |f (x) L| < . Suppose that
0 < a x < . Then 0 < a x = |x a| < |f (x) L| < . This shows
that L = limxa f (x). Likewise, one can show that L = limxa+ f (x).
Conversely, suppose limxa f (x) = limxa+ f (x) = L. Let  > 0 be given.
Then there exist 1 > 0 and 2 > 0 such that 0 < a x < 1 |f (x) L| < 
and 0 < x a < 2 |f (x) L| < . Let = min{1 , 2 }. Suppose
0 < |x a| < . Since |x a| = (x a) we must have |f (x) L| < . This
shows that limxa f (x) = L

Exercise 8.9
Using  and , what does it mean that limxa f (x) 6= L?

Solution. If lim x af (x) 6= L then there is an  > 0 such that for all > 0
there is an x in the domain of f such that 0 < |xa| < but |f (x)L| 

47
Solutions to Section 9
Exercise 9.1
Suppose that limxa f (x) = L1 and limxa g(x) = L2 . Show that

lim [f (x) g(x)] = L1 L2 .

xa

Solution.
Let  > 0 be arbitrary. Since limxa f (x) = L1 , we can find 1 > 0 such that

0 < |x a| < 1 = |f (x) L1 | < .
2
Similarly, since limxa g(x) = L2 , we can find 2 > 0 such that

0 < |x a| < 2 = |g(x) L2 | < .
2
Let = min{1 , 2 }. Notice that 1 and that 2 . Thus, if 0 < |xa| <
then

|(f (x) + g(x)) (L1 + L2 )| =|(f (x) L1 ) + (g(x) L2 )|

|f (x) L1 | + |g(x) L2 |
 
< + =
2 2
A similar argument holds for f (x) g(x)

Exercise 9.2
Suppose that limxa f (x) = L1 and limxa g(x) = L2 . Show the following:
(a) There is a 1 > 0 such that

0 < |x a| < 1 = |f (x)| < 1 + |L1 |.

Hint: Notice that f (x) = (f (x) L1 ) + L1 .

(b) Given  > 0, there is a 2 > 0 such that

0 < |x a| < 2 = |f (x) L1 | < .
2(1 + |L2 |)

48
Solution.
(a) Let  = 1. Since limxa f (x) = L1 , we can find 1 > 0 such that

0 < |x a| < 1 = |f (x) L1 | < 1.

But then

|f (x)| = |(f (x) L1 ) + L1 | |f (x) L1 | + |L1 | < 1 + |L1 |.


(b) Since 2(1+|L2 |)
> 0 and limxa f (x) = L1 , we can find 2 > 0 such that


0 < |x a| < 2 = |f (x) L1 | <
2(1 + |L2 |)

Exercise 9.3
Suppose that limxa f (x) = L1 and limxa g(x) = L2 .
(a) Show that f (x)g(x) L1 L2 = f (x)(g(x) L2 ) + L2 (f (x) L1 ).
(b) Show that |f (x)g(x) L1 L2 | |f (x)||g(x) L2 | + |L2 ||f (x) L1 |.
(c) Show that limxa f (x)g(x) = L1 L2 . Hint: Use the previous exercise.

Solution.
(a) We have

f (x)(g(x) L2 ) + L2 (f (x) L1 ) =f (x)g(x) L2 f (x) + L2 f (x) L1 L2

=f (x)g(x) L1 L2 .

(b) By (a) and Exercise 1.7/Exercise1.17, we have

|f (x)g(x) L1 L2 | =|f (x)(g(x) L2 ) + L2 (f (x) L1 )|

|f (x)(g(x) L2 )| + |L2 (f (x) L1 )|
=|f (x)||(g(x) L2 )| + |L2 ||(f (x) L1 )|

(c) Let  > 0 be arbitrary. By Exercise 9.2(a), there is 1 > 0 such that

0 < |x a| < 1 = |f (x)| < 1 + |L1 |.

Also, by Exercise 9.2(b), there exist 2 > 0 and 3 > 0 such that

0 < |x a| < 2 = |f (x) L1 | <
2(1 + |L2 |)

49
and

0 < |x a| < 3 = |g(x) L2 | < .
2(1 + |L1 |)
Let = min{1 , 2 , 3 }. Notice that 1 , 2 , and 3 . Suppose that
0 < |x a| < . Using (b) and the above inequalities we find

|f (x)g(x) L1 L2 | |f (x)||g(x) L2 | + |L2 ||f (x) L1 |

 
<(1 + |L1 |) + |L2 |
2(1 + |L1 |) 2(1 + |L2 |)
 
<(1 + |L1 |) + (1 + |L2 |)|
2(1 + |L1 |) 2(1 + |L2 |)
 
= + =
2 2
This establishes the required result

Exercise 9.4
(a) Suppose that |f (x)| M for all x in its domain and limxa g(x) = 0.
Show that
lim f (x)g(x) = 0.
xa

Hint: Recall Exercise 4.5

1


(b) Show that limx0 x sin x
= 0.

Solution.
(a) Let  > 0 be arbitrary. Since limxa g(x) = 0, there is a positive number
such that if 0 < |x a| < then |g(x) 0| = |g(x)| < M . Thus, for
0 < |x a| < we have

|f (x)g(x) 0| = |f (x)g(x)| = |f (x)||g(x)| < M = .
M
This shows that limxa f (x)g(x) =
0.
(b)
 Let f (x) = x and g(x) = sin x1 . Note that limx0 f (x) = 0 and |g(x)| =
sin 1  1. It follows from (a) that


x
 
1
lim x sin = lim f (x)g(x) = 0
x0 x x0

50
Exercise 9.5
Suppose that limxa f (x) = L with L 6= 0. Show that there exists a > 0
such that
|L|
0 < |x a| < = |f (x)| > > 0.
2
Hint: Recall the solution to Exercise 4.7

Solution.
Let  = |L|
2
> 0. Then there exists > 0 such that |f (x) L| < |L|
2
whenever
0 < |xa| < . But by Exercise 1.18, we have |L||f (x)| < 2 or |f (x)| > |L|
|L|
2
whenever 0 < |x a| <

Exercise 9.6
Let g(x) be a function with the following conditions:
(1) g(x) 6= 0 for all x in the domain of g.
(2) limxa g(x) = L2 , with L2 6= 0.
(a) Show that there is a 1 > 0 such that if 0 < |x a| < 1 then
 
 1 1  2
 g(x) L2  < |L2 |2 |g(x) L2 |.

 

(b) Let  > 0 be arbitrary. Show that there is 2 > 0 such that if 0 < |xa| <
2 then
|L2 |2
|g(x) L2 | < .
2
(c) Using (a) and (b), show that
1 1
lim = .
xa g(x) L2

Hint: Recall Exercise 4.8

Solution.
|L2 |
(a) From Exercise 9.5, we can find 1 > 0 such that |g(x)| > 2
whenever
0 < |x a| < 1 . Thus, for 0 < |x a| < 1 we obtain
 
 1 1  |g(x) L2 | 2
 g(x) L2  = |g(x)||L2 | < |L2 |2 |g(x) L2 |.


51
(b) Let  > 0 be arbitrary. Since limxa g(x) = L2 , we can find a positive
number 2 such that if 0 < |x a| < 2 then

|L2 |2
|g(x) L2 | < .
2
(c) Let  > 0 be arbitrary. Let = min{1 , 2 }. Notice that 1 and
2 . Now, if 0 < |x a| < then

2 |L2 |2
 
 1 1  |g(x) L2 | 2
 = < |g(x) L 2 | <  = .
 g(x) L2  |g(x)||L2 | |L2 |2 |L2 |2 2

This establishes the required result

Exercise 9.7
Show that if limxa f (x) = L1 and limxa g(x) = L2 where g(x) 6= 0 in its
domain and L2 6= 0 then
f (x) L1
lim = .
xa g(x) L2
Hint: Recall Exercise 4.10.

Solution.
Using Exercise 9.6 and Exercise 9.3 we can write
 
f (x) 1 1 L1
lim = lim f (x) = L1 =
xa g(x) xa g(x) L2 L2

Exercise 9.8
Let f (x) and g(x) be two functions with a common domain D and a a point
in D. Suppose that f (x) g(x) for all x in D. Show that if limxa f (x) = L1
and limxa g(x) = L2 then L1 L2 . Hint: Recall Exercise 4.11

Solution.
Since limxa f (x) = L1 , there exists a positive number 1 such that |f (x)
L1 | <  whenever 0 < |x a| < 1 . By Exercise 1.14, this is equivalent
to L1  < f (x) < L1 +  whenever 0 < |x a| < 1 . Similarly, since
limxa g(x) = L2 , there exists a positive number 2 such that |g(x) L2 | < 
whenever 0 < |x a| < 2 . By Exercise 1.14, this is equivalent to L2  <

52
g(x) < L2 +  whenever 0 < |x a| < 2 .
Suppose that L1 > L2 . Let  = L1 L
2
2
. We have

L 1 L2 L1 + L2
g(x) <L2 +  = L2 + =
2 2
L 1 L2
=L1 = L1  < f (x)
2
This contradicts the fact that f (x) g(x) for all x in D. Hence, we conclude
that L1 L2

Exercise 9.9
Let D be the domain of a function f (x). Suppose that f (x) 0 for all x in
D and limxa f (x) = L with L > 0.
(a) Show that
p f (x) L
f (x) L = p .
f (x) + L

(b) Let  > 0. Show that there exists > 0 such that |f (x) L| <  L
whenever 0 < |x a| < .
(c) Show that p
lim f (x) = L.
xa

Solution.
(a) We have
p

p
p ( f (x) L)( f (x) + L) f (x) L
f (x) L = p =p .
f (x) + L f (x) + L

(b) Since
lim xa f (x) = L and  L > 0 there exists > 0 such that |f (x)
L| <  L whenever 0 < |x a| < .
(c) For 0 < |x a| < we have
 
p  f (x) L  |f (x) L|  L
| f (x) L| =  p  < = .
 
 f (x) + L  L L

This establishes a proof of the required result

53
Exercise 9.10 (Squeeze Rule)
Let f (x), g(x) and h(x) be three functions with common domain D and a be
a point in D. Suppose that
(1) g(x) f (x) h(x) for all x in D.
(2) limxa g(x) = limxa h(x) = L.
Show that limxa f (x) = L. Hint: Recall Exercise 3.11

Solution.
Let  > 0. From (2) we can find 1 > 0 such that if 0 < |x a| < 1 then

|g(x) L| <  L  < g(x) < L + .

Similarly, we can find 2 > 0 such that if 0 < |x a| < 2 then

|h(x) L| <  L  < h(x) < L + .

Let = min{1 , 2 }. Suppose 0 < |x a| < . Then L  < g(x) f (x)

h(x) < L + . That is L  < f (x) < L +  whenever 0 < |x a| < . This
implies that limxa f (x) = L

Exercise 9.11
Consider the following figure.

where 0 < x < 2 .

(a) Using geometry, establish the inequality

0 < sin x < x.

Hint: The area of a circular sector with radius r and central angle is given
by the formula 21 r2 .

54
(b) Show that limx0+ sin x = 0.
(c) Show that limx0 sin x = 0. Thus, we conclude that limx0 sin x = 0.
Hint: Recall that the sine function is an odd function.
(d) Show that limx0 cos x = 1. Hint: cos2 x + sin2 x = 1.
(e) Using geometry, establish the double inequality
sin x cos x x tan x
< < .
2 2 2
(f) Using (a) show that
sin x 1
cos x < < .
x cos x
(g) Show that
sin x
lim+ = 1.
x0 x
(h) Show that for 2 < x < 0 we have also
sin x
lim = 1.
x0 x
Solution.
(a) The area of the triangle OBC is smaller than the circular area. Thus
sin x x
0< <
2 2
or
0 < sin x < x.
(b) This follows from (a) and the squeeze rule.
(c) For x < 0 we can write

lim sin x = lim sin (x) = lim + sin (x) = 0.

x0 x0 x0
p
(d) Since cos x 0 for 2 x 2 we have cos x = 1 sin2 x. Thus,
using Exercise 9.9, we find
p q
lim cos x = lim 1 sin2 x = lim (1 sin2 x) = 1 0 = 1.
x0 x0 x0

sin x cos x
(e) The area of the triangle OAB is 2
. The area of the triangle OCD
is tan2 x . From the graph we see that

55
 Area of triangle OAB < Area of circular sector < Area of triangle OCD.
Thus,
sin x cos x x tan x
< < .
2 2 2
(f) Since 0 < x < 2 , sin x > 0. Simple algebra leads to
sin x 1
cos x < < .
x cos x
(g) Notice that limx0 cos x = 1 and limx0 cos1 x = 1. Now the result follows
by applying the squeeze rule.
(h) Since sin x = sin (x) and 0 < x < 2 for 2 < x < 0, we can use
(g) to obtain
sin x sin (x)
lim = lim + =1
x0 x x0 x
Exercise 9.12
Find each of

the following

limits:
x2 +32 x
(1) limx1 x2 1
.
x2
(2) limx2 |x2 5x+6| .
Solution.
(a) We have

x2 + 3 2 x ( x2 + 3 2 x)( x2 + 3 + 2 x
lim = lim
x1 x2 1 x1 ( x2 + 3 + 2 x)(x2 1)
x2 4x + 3
= lim
x1 ( x2 + 3 + 2 x)(x2 1)

(x 1)(x 3)
= lim
x1 ( x2 + 3 + 2 x)(x 1)(x + 1)

x3 1
= lim =
x1 ( x2 + 3 + 2 x)(x + 1) 4
(2) We have
x2 x2
lim = lim
x2 |x2 5x + 6| x2 |(x 2)(x 3)|
x2
= lim
x2 (x 2)(x 3)
1
= lim = 1
x2 x 3

56
Exercise 9.13
2
Find limx xx2 +x
x
by using the change of variable u = x1 .

Solution.
1
By letting u = x
we find

x2 + x 1+u
lim = lim =1
x x2 x u0 1 u

Exercise 9.14

Find limx0 3 x sin x1 .

Solution.
We have for x > 0
1
3 x 3 x sin 3 x
x
and for x < 0
1
3 x 3 x sin 3 x
x

By the squeeze rule we conclude that limx0 3 x sin x1 = 0

Exercise 9.15
Find limx0 x2 tan x.

Solution.
For 4 x 4 we have 1 tan x 1. Thus, x2 x2 tan x x2 . By
the squeeze rule we conclude that limx0 x2 tan x = 0

Exercise 9.16
Let n be a positive integer. Prove that limxa [f (x)]n = [limxa f (x)]n .

Solution.
Follows by a simple induction on n in Exercise 9.3

57
Solutions to Section 10
Exercise 10.1
Suppose that limxa f (x) = L, where a is in the domain of f. Let {an } n=1
be a sequence whose terms belong to the domain of f and are different from
a and suppose that limn an = a.
(a) Let  > 0 be arbitrary. Show that there exist a positive integer N and
a positive number such that for n N we have |an a| < and for
0 < |x a| < we have |f (x) L| < .
(b) Use (a) to conclude that for a given  > 0 there is a positive integer N
such that if n N then |f (an ) L| < . That is, limn f (an ) = L.

Solution.
(a) Let  > 0 be given. Since limxa f (x) = L, we can find a positive constant
such that if 0 < |x a| < then |f (x) L| < . But for > 0 we can
find a positive integer N such that for n N we have |an a| < due to
the fact that limn an = a. Since an 6= a for all n 1 we conclude that
0 < |an a| < for all n N.
(b) Using (a), for a given  > 0 we can find a positive integer N such that
if n N then 0 < |an a| < which implies that |f (an ) L| < . This
establishes that
lim f (an ) = L
n

Exercise 10.2
Let {an }n=1 be an arbitrary sequence of terms in the domain of f with an 6= a
for all n 1. Suppose that if limn an = a then limn f (an ) = L. Clearly,
either

limxa f (x) = L or limxa f (x) 6= L.

(a) Suppose first that limxa f (x) 6= L. Show that there is an  > 0 and a
sequence {an } 1
n=1 of terms in the domain of f such that 0 < |an a| < n and
|f (an ) L| .
(b) Use the squeeze rule to show that limn |an a| = 0.
(c) Use the fact that |a| a |a| for any number a and the squeeze rule
to show that limn (an a) = 0.
(d) Use Exercise 4.1 to show that limn an = a.
(e) Using (a), (d), the given hypothesis and Exercise 4.11, show that  0.

58
Thus, this contradiction shows that limxa f (x) 6= L cannot happen. We
conclude that
lim f (x) = L
xa

Solution.
(a) Since limxa f (x) 6= L, there is  > 0 such that for every n 1, there is
an in the domain of f such that 0 < |an a| < n1 but |f (an ) L| .
(b) Let bn = 0, cn = |an a|, and dn = n1 . Then we have bn < cn < dn . Since
limn bn = limn n1 = 0. By the squeeze rule we obtain limn |an a| =
0.
(c) Since |an a| an a |an a|, by the squeeze rule we find limn (an
a) = 0.
(d) Since an = (an a) + a, we can use Exercise 4.1 to obtain limn an =
limn (an a) + limn a = 0 + a = a.
(e) According to (a) and (d), we can find a sequence {an } n=1 of terms in the
domain of f such that an 6= a for all n 1, |f (an ) L| . By hypothesis,
limn f (an ) = L. By Exercise 4.11,  0, a contradiction. Hence, we must
have
lim f (x) = L
xa

Exercise 10.3
Let f be a function with domain D and a be a point in D. Suppose that f
satisfies the following Property:

(P) If {an }
n=1 , with an in D, an 6= a for all n 1 and limn an = a then
{f (an )}
n=1 is a Cauchy sequence.

(a) Let {an }

n=1 be a sequence of elements of D such that an 6= a for all n 1
and limn an = a. Show that the sequence {f (an )} n=1 is convergent. Call
the limit L. Hint: See Exercise 7.7
(b) Let {bn }
n=1 be a sequence of elements of D such that bn 6= a for all n 1
and limn bn = a. Show that the sequence {f (bn )} n=1 converges to L .
0

Solution.
(a) By Property (P) the sequence {f (an )} n=1 is a Cauchy sequence. By Ex-

erccise 7.7, the sequence {f (an )}n=1 converges, say with limit L.
(b) By an argument similar to (a), there is a number L0 such that limn f (bn ) =
L0

59
Exercise 10.4
Let {an }
n=1 and {bn }n=1 be the two sequences of the previous exercise. Define
the sequence
{cn } = {b1 , a1 , b2 , a2 , b3 , a3 , }.
That is, cn = ak if n = 2k and cn = bk if n = 2k + 1 where k 0.
(a) Show that for all n we have cn D and cn 6= a.
(b) Let  > 0. Show that there exist positive integers N1 and N2 such that if
n N1 then |an a| <  and if n N2 then |bn a| < .
(c) Let N = 2N1 + 2N2 + 1. Show that if n N then |cn a| < . Hence,
limn cn = a. Hint: Consider the cases n = 2k or n = 2k + 1.
(d) Show that limn f (cn ) = L00 for some number L00 .

Solution.
(a) Since cn is either an or bn and both belong to D and are different from
a, we conclude that cn belongs to D and cn 6= a for all n 1.
(b) Let  > 0 be arbitrary. Since limn an = a and limn bn = a we can
find positive integers N1 and N2 such that if n N1 we have |an a| < 
and if n N2 we have |bn a| < .
(c) Let N = 2N1 + 2N2 + 1. Let n N. If n = 2k + 1 then 2k + 1 N
2N2 + 1 k N2 so that |cn a| = |bk a| < . If n = 2k N 2N1 then
k N1 and in this case |cn a| = |ak a| < . Thus, whether n is even or
odd we have |cn a| < . It follows that

lim cn = a.
n

(d) With an argument similar to Exercise 10.2, we conclude that limn f (cn ) =
L00 for some number L00

Exercise 10.5
Let {an }
n=1 , {bn }n=1 , and {cn }n=1 be as in the previous exercise.
(a) Compare {an }
n=1 and {cn }n=1 .
(b) Let  > 0 be arbitrary. Show that there is a positive integer N such that
if n N then |f (cn ) L00 | < .
(c) Let N1 be a positive integer such that N1 N2 . Show that if n N1 then
|f (an ) L00 | < . Hence, limn f (an ) = L00 .
(d) Show that limn f (bn ) = L00 . Thus, by Exercise 3.6, we must have
L = L0 = L00 .

60
Solution.
(a) First note that {an }
n=1 = {c2n }n=1 .
(b) Let  > 0 be arbitrary. Since limn f (cn ) = L00 , there is a positive
integer N such that if n N we have |f (cn ) L00 | < .
(c) Let N1 be a positive integer such that N1 N2 . For n N1 2n N
|f (an ) L00 | = |f (c2n ) L00 | < . Hence, limn f (an ) = L00 . by Exercise 3.6,
we obtain L = L00 .
(d) First note that {bn }
n=1 = {c2n1 }n=1 . Let  > 0 be arbitrary. Since
limn f (cn ) = L00 , there is a positive integer N such that if n N we
have |f (cn ) L00 | < . Let N2 be a positive integer such that N2 N 2+1 .
For n N2 2n 1 N |f (bn ) L00 | = |f (c2n1 ) L00 | < . Hence,
limn f (bn ) = L00 . by Exercise 3.6, we obtain L = L00

Exercise 10.6
Prove that if a function f satisfies property (P) then limxa f (x) exists. Hint:
Use Exercise 10.2.

Solution.
This follows from Exercise 10.2 and Exercise 10.5

Exercise 10.7
Consider the function f : R R defined by

sin x1 if x 6= 0

f (x) =
1 if x = 0

Let {an }
n=1 and {bn }n=1 be the two sequences defined by an = 2n and
1

bn = (2n+1 1 ) . Clearly, an , bn 6= 0 for all n N, an 0 and bn 0. Show

2
that limx0 f (x) does not exist.

Solution.
Since limn f (an ) = 0 and limn f (bn ) = 1, by Exercise 10.1, limx0 f (x)
does not exist

Exercise 10.8
Let {an }
n=1 be a sequence such that an 6= 2 for all n N and limn an = 2.
2 4
(a) Find limn aann 2 = 4.
x2 4
(b) Find limx2 x2 .

61
Solution.
(a) We have
a2n 4
lim = lim (an + 2) = 4.
n an 2 n

(b) By Exercise 10.2, we have

x2 4
lim =4
x2 x 2

Exercise 10.9
Consider the floor function f : [0, 1] R given by f (x) = bxc, where bxc
denote the largest integer less than or equal to x. Find limx1 bxc using
sequences.

Solution.
Let {an }
n=1 [0, 1] such that an 6= 1 for all n N and limn an = 1. Since
an (0, 1) we have f (an ) = 0 for all n N. Thus, limn f (an ) = 0. We
conclude that limx1 f (x) = 0

Exercise 10.10
Consider the floor function f : R R given by f (x) = bxc, where bxc denote
the largest integer less than or equal to x.
(a) Let an = 1 n1 and bn = 1 + n1 for all n N. Find limn f (an ) and
limn f (bn ).
(b) Does limx1 bxc exist?

Solution.
(a) bf (an )c = 0 for all n N so limn f (an ) = 0. Likewise, bf (bn )c = 1 for
all n N so limn f (bn ) = 1.
(b) By Exercise ??, limx1 bxc does not exist

62
Solutions to Section 11
Exercise 11.1
Show that the function f (x) = x2 is continuous at x = 0.

Solution.
Let  > 0 be arbitrary. Let 0 < < . Then for any x such that |x 0| <
we have |x2 0| = |x|2 < 2 < . This shows that f (x) = x2 is continuous at
x=0

Exercise 11.2
Show that f is continuous at x = a if and only if limxa f (x) = f (a).

Solution.
Suppose first that f is continuous at x = a. Let  > 0 be arbitrary. Then,
there is a > 0 such that for all x in D, if |x a| < then |f (x) f (a)| < .
In particular, for all x in D such that x 6= a if 0 < |x a| < then
|f (x) f (a)| < . But this is the same as

lim f (x) = f (a).

xa

Conversely, suppose that limxa f (x) = f (a). Let  > 0 be given. There is
a > 0 such that for all x in D, if 0 < |x a| < then |f (x) f (a)| < .
Since |a a| = 0 < and |f (a) f (a)| = 0 <  then for all |x a| < we
have |f (x) f (a)| < . This shows that f is continuous at x = a

Exercise 11.3
Consider the function
x2 4

x2
if x 6= 2
f (x) =
0 if x = 2

Show that f is discontinuous at x = 2.

Solution.
2 4
We have limx2 xx2 = limx2 (x2)(x+2)
x2
= limx2 (x + 2) = 4 6= f (2). By
Exercise 11.2, the function is discontinuous at 2

63
Exercise 11.4
Suppose that f is discontinuous at x = a.
(a) Show that there is and  > 0 and a sequence {an }n=1 of elements in D
1
such that 0 |an a| < n and |f (an ) f (a)| .
(b) Show that limn |an a| = 0.
(c) Show that limn an = a

Solution.
(a) By Definition 11, there is an  > 0 such that for each n = n1 we can find
an in D with 0 |an a| < n1 and |f (an ) f (a)| .
(b) Using the Squeeze rule we conclude that limn |an a| = 0.
(c) Since |an a| an a |an a|, by the Squeeze rule we conclude that
limn (an a) = 0. But limn an = limn (an a) + a = limn (an
a) + limn a = 0 + a = a

Exercise 11.5
Suppose that f is continuous at x = a. Let {an }n=1 be a sequence of elements
in D converging to a.
(a) Let  > 0 be given. Show that there is a > 0 such that for any x in D
such that |x a| < we have |f (x) f (a)| < .
(b) With the  and as in (a), show that there is a positive integer N such
that if n N then |an a| < .
(c) Conclude that limn f (an ) = f (a).

Solution.
(a) Let  > 0 be given. Since f is continuous at x = a, there is a > 0 such
that for any x in D with |x a| < we have |f (x) f (a)| < .
(b) Since limn an = a, for the > 0 we can find a positive integer N such
that if n N then |an a| < .
(c) Using (a) and (b) we see that for any  > 0 we can find a positive integer N
such that if n N we have |an a| < which implies that |f (an )f (a)| < .
This establishes the result

lim f (an ) = f (a)

n

Exercise 11.6
Suppose that for any sequence {an }n=1 of elements in D that converges to a,
the sequence {f (an )}
n=1 converges to f (a). Then either f is continuous at a

64
or f is discontinuous at a.
(a) Suppose that f is discontinuous at a. Show that there is an  > 0 and
a sequence {an } n=1 of elements in D such that limn an = a and |f (an )
f (a)|  for all n 1.
(b) Show that limn f (an ) = f (a).
(c) Show that by (a) and (b) we conclude that  0, a contradiction. Thus,
f must be continuous at x = a.

Solution.
(a) This follows from Exercise ??.
(b) By hypothesis, since limn an = a we must have limn f (an ) = f (a)
or limn (f (an ) f (a)) = limn f (an ) limn f (a) = f (a) f (a) = 0.
(c) Since 0 (f (an ) f (a)) |f (an ) f (a)| (f (an ) f (a)) so that by
the Squeeze rule limn |f (an ) f (a)| = 0. Since, 0 <  |f (an ) f (a)| for
all n 1, we can apply the Squeeze rule and get  0, a contradiction. We
conclude that f must be continuous at x = a

Exercise 11.7
Consider the function 
1 if x 0
f (x) =
0 if x < 0
(a) Let an = n1 . Find limn an and limn f (an ).
(b) Is f continuous at x = 0?

Solution.
(a) We have limn an = 0 and limn f (an ) = 0.
(b) From (a) and Exercise 11.5, the function is discontinuous at x = 0

Exercise 11.8
Give an example of a continuous function f : R R and a sequence {an }
n=1
such that limn f (an ) exists, but limn an does not exist.

Solution.
Let an = (1)n . We know that this sequence is divergent so that limn an
does not exist. However, f (an ) = 1 for all n N so that limn f (an ) = 1

65
Exercise 11.9
Determine the values of a and b that makes the function f continuous every-
where. sin x
2 x if x < 0
f (x) = a if x = 0
b cos x if x > 0

Solution.
We must have limx0 f (x) = limx0+ f (x) = f (0) = a. But limx0 f (x) =
2 so that a = 2. Also, limx0+ f (x) = b = 2

Exercise 11.10
Using the - definition of continuity show that f (x) = x3 is continuous at
x = 1. Hint: x3 1 = (x 1)(x2 + x + 1).

Solution.
We first note that |x3 1| = |(x1)(x2 +x+1)| = |x1||x2 +x+1| 3|x1|.
Let  > 0 be given. Let < 5 . Then |x 1| < |f (x) 1| < . This shows
that f (x) is continuous at x = 1

Exercise 11.11
Consider the function f (x) = cos x1 .


1
(a) Let an = 2n and bn = (n+11 ) . Find limn an , limn bn , limn f (an ),
2
and limn f (bn ).
(b) Is f continuous at x = 0?

Solution.
(a) we have limn an
= limn bn = 0, limn f (an ) = 1, and limn f (bn ) =
0. Thus, limx0 cos x1 does not exist and therefore the function is not con-
tinuous at x = 0

Exercise 11.12
Consider the function
1


x sin x
if x 6= 0
f (x) =
0 if x = 0
Show that this function is continuous at x = 0 by using the - definition.

Solution.  
1


Let  > 0 be given. Choose = . Then |x| < x sin x
0 |x| < 

66
Exercise 11.13
Prove that if f is continuous at x = a so does |f |. Hint: Exercise 1.23.

Solution.
Let  > 0 be given. Then there is a > 0 such that |x a| < delta
|f (x) f (a)| < . But ||f (x)| |f (a)|| |f (x) f (a)| < . This shows that
|f | is continuous at x = a

Exercise 11.14
Suppose f, g : R R are continuous on R. Suppose h : R R satisfies
f (x) h(x) g(x) for all x R. If f (c) = g(c), prove that h is continuous
at c.

Solution.
Let {an }
n=1 be any sequence in R that converges to c. We must show that the
sequence {h(an )}n=1 converges to h(c). By the continuity of f at c, we know
that {f (an )}n=1 converges to f (c), and similarly we know that {g(an )}

n=1
converges to g(c). We also have the inequality f (an ) h(an ) g(an ). Since
limn f (an ) = f (c) = g(c) = limn g(an ), it follows from the Squeeze
theorem that {h(an )} n=1 converges, and limn h(an ) = f (c) = h(c). Thus
h is continuous at c

Exercise 11.15
Let f : [0, ) R be defined by f (x) = x. Show that f is continuous on
[0, ).

Solution.
First we show that f is right continuous
at x = 0. Let  > 0 be given. Choose
= 2 . Then if 0 < x < we have | x 0| = x < .
Now let c > 0. Let  > 0 be given. Let <  c. Then
 
 ( x c)( x c) 
|f (x) f (c)| = 
 
x+ c 
|x c|
=
x+ c
1
< c = 
c

67
Solutions to Section 12
Exercise 12.1
Let f (x) and g(x) be two functions with common domain D. Suppose that
f and g are continuous at a point a in D. Show the following properties:
(i) f g is continuous at a.
(ii) f g is continuous at a.
(iii) fg is continuous at a provided that g(a) 6= 0.

Solution.
(i) By Exercise 9.1, we have

lim (f g)(x) = lim f (x) lim g(x) = f (a) g(a) = (f g)(a).

xa xa xa

By Exercise 11.2, (f g) is continuous at a.

(ii) By Exercise 9.3, we have

lim (f g)(x) = lim f (x) lim g(x) = f (a) g(a) = (f g)(a).

xa xa xa

By Exercise 11.2, (f g) is contiuous at a.

(iii) By Exercise 9.7, we have
   
f limxa f (x) f (a) f
lim (x) = = = (a).
xa g limxa g(x) g(a) g
f
By Exercise 11.2, g
is contiuous at a

Exercise 12.2
Let f be continuous at a point a in its domain with f (a) 6= 0. Show that
there exists a > 0 such that
|f (a)|
|x a| < = |f (x)| > 2
.

That is, there is an open interval centered at a where the function is always
different from zero there. Hint: Look at Exercise 4.7

Solution.
Let  = |f (a)|
2
> 0. By the definition of continuity, there is a > 0 such that
|f (a)|
|x a| < = |f (x) f (a)| < 2
.

68
 |f (a)| |f (a)|
By Exercise 1.18, we have |f (a)| |f (x)| < 2
or |f (x)| > 2

Exercise 12.3
Let f : D R and g : D0 R with the range of f contained in D0 . Thus,
g f : D R is a function with domain D. Suppose that f is continuous at
a and g is continuous at f (a).
(a) Let  > 0 be given. Show that there is a 0 > 0 such that for all y in D0
satisfying |y f (a)| < 0 we have |g(y) g(f (a))| < .
(b) Show that there is a 00 > 0 such that if |xa| < 00 then |f (x)f (a)| < 0 .
(c) Show that there is a > 0 such that if |xa| < then |g(f (x))g(f (a))| <
. In other words, the composite function g(f (x)) is continuous at a. Hence,
the composition of two continuous functions is a continuous function.

Solution.
(a) Let  > 0 be given. Since g is continuous at f (a), that is

lim g(y) = g(f (a))

yf (a)

there is a 0 > 0 such that if |y f (a)| < 0 then |g(y) g(f (a))| < , where
y is in D0 .
(b) Now, since f is continuous at a, there is a 00 > 0 such that if |x a| < 00
then |f (x) f (a)| < 0 .
(c) Let = min{ 0 , 00 }. If |x a| < then |x a| < 00 which implies that
|f (x) f (a)| < 0 . Letting y = f (x) we find that |y f (a)| < 0 . But then
|g(y) g(f (a))| = |g(f (x)) g(f (a))| < . This establishes the fact that
g(f (x)) is continuous at a

Exercise 12.4
In Exercise 9.11, we established that limx0 sin x = 0 = sin 0. That is, the
sine function is continuous at 0.
(a) Using the trigonometric identity

sin (a + b) = sin a cos b + cos a sin b

show that the sine function is continuous at every number a. Hint: Use the
substitution u = x a and note that u 0 as x a.
(b) Show that the cosine function is continuous for every number a. Hint:



Note that cos x = sin 2 x and use Exercise 12.3.

69
Solution.
(a) Letting u = x a we have

lim sin x = lim sin (u + a)

xa u0
= lim (sin u cos a + cos u sin a)
u0
= lim sin u cos a + lim cos u sin a
u0 u0
=0 + (1) sin a = sin a

(b) Let f
(x) = 2 x, g(x) = sin x, and h(x) = cos x. Then h(x) =
sin 2 x = g(f (x)). Since f and g are continuous for every real number,
by the previous exercise, h is continuous for every real number

Exercise 12.5
Suppose that f : R R is continuous such that f (x) = 0 for all x Q.
Prove that f (x) = 0 for all x R. Hint: Exercise 3.21

Solution.
Let a be an irrational number. By Exercise 3.21, there is a sequence of
rational numbers {an } n=1 that converges to a. By continuity, we must have
f (a) = limn f (an ) = 0. Thus, f (x) = 0 for all x R

Exercise 12.6
Consider the function 
x if x Q
f (x) =
0 if x 6 Q
(a) Prove that f is continuous at x = 0.
(b) Let a 6= 0. Prove that f is discontinuous at x = a.

Solution.
(a) Let  > 0 be given. Choose = . If x is rational such that |x| < then
|f (x)f (0)| = |x| < . If x is irrational such that |x| < then |f (x)f (0)| =
0 < . Hence, f is continuous at x = 0.
(b) Let a 6= 0. We can find a sequence of rationals {an } n=1 that converges to
a. In this case, limn f (an ) = limn an = a. Also, we can find a sequence
of irrationals {bn }
n=1 that converges to a. In this case, limn f (bn ) = 0.
Thus, limxa f (x) does not exist and so the function is discontinuous at
x=a

70
Exercise 12.7
Suppose f, g : R R are continuous functions and f (x) = g(x) for every
x Q. Show that f (x) = g(x) for every x R.

Solution.
Apply Exercise 12.5 to the function h(x) = f (x) g(x)

Exercise 12.8
Use continuity to evaluate limx sin (x + sin x).

Solution.
We have limx sin (x + sin x) = sin ( + 0) = 0

Exercise 12.9
Give an example of two functions f and g that are not continuous on the
interval (0, 1) but their sum f + g is continuous on (0, 1).

Solution.
Let
1 if 0 < x < 12

f (x) =
2 if 12 x < 1
and
2 if 0 < x < 21

g(x) =
1 if 12 x < 1
Thus, f and g are discontinuous at x = 21 . However, (f + g)(x) = 3 for all
x (0, 1) which is a continuous function

Exercise 12.10
Let f : R R be a continuous function that satisfies f (x + y) = f (x) + f (y)
for all x, y R.
(a) Show that f (0) = 0 and f (n) = an and f (n) = a(n) for all n N
where a = f (1).

(b) Show f m n
= a mn
where m and n are integers with n 6= 0. That is,
f (x) = ax for all x Q.
(c) Show that f (x) = ax for all x R. Hint: Exercise 12.5 applied to the
function g(x) = f (x) ax.

71
Solution.
(a) We have f (0) + f (0) = f (0 + 0) = f (0). Subtract f (0) from both sides
to obtain f (0) = 0. Let a = f (1) and n N. Write n = 1 + 1 + + 1. Then
f (n) = f (1 + 1 + + 1) = f (1) + f (1) + + f (1) = nf (1) = an. Also,
f (n) + f (n) = f (n n) = f (0) = 0 so that f (n) = f (n)
= a(n).1

m
(b) Let
m and n be
integers such that n 6= 0. Then f n
= f m n =
mf n . Since nf n = f n = f (1) = a we obtain f m
1 1 n
= am


n n
. Thus,
f (x) = ax for all x Q.
(c) Let g(x) = f (x) ax. g is continuous on R and g(x) = 0 for all x Q. by
Exercise 12.5, we conclude that g(x) = 0 for all x R. That is, f (x) = ax
for all x R

Exercise 12.11
Prove that if f is continuous on [a, b], then either f (x) = 0 for some x [a, b],
or there is a number  > 0 such that |f (x)|  for all x [a, b].

Solution.
Suppose that for every  > 0 there is an x [a, b] such that |f (x)| < . By let-
ting  = n1 we find a sequence {an } 1
n=1 such that |f (an )| < n . Since {an }n=1 is
bounded (since all the terms are in [a, b]) we can find a subsequence {ank } k=1
of {an } 1
n=1 that converges to some number x [a, b]. Since |f (ank )| < nk and
f is continuous we conclude that f (x) = limn f (ank ) = 0. Thus either
there is some  > 0 with |f (x)|  for all x [a, b], or there is some x [a, b]
with f (x) = 0

72
Solutions to Section 13
Exercise 13.1
Show that the function f (x) = x is uniformly continuous.

Solution.
Let  > 0 be given. Let = . Then for all x1 and x2 , if |x1 x2 | < then
|f (x1 ) f (x2 )| = |x1 x2 | < = . Hence, the given function is uniformly
continuous

Exercise 13.2
Consider the function f (x) = x1 on the set x > 0. Let > 0 be any number
and define = min{2, }. Then 2 and . Let x1 = 3 > 0 and
x2 = 6 > 0.
(a) Show that |x1 x2 | < but |f (x1 ) f (x2 )| 32 .
(b) Conclude from (a) that f is not uniformly continuous on the interval
0 < x < .

Solution.  
(a) We have |x 1 x 2 | =   = < . Moreover, |f (x1 ) f (x2 )| =
3  3 3 6 6 6
6
 = . 3
2
(b) By letting  = 23 there is no > 0 where the definition of uniform
continuity is satisfied. Hence, f (x) = x1 is not uniformly continuous on
(0, )

Exercise 13.3
(a) Show that if f is uniformly continuous on D then f is continuous at every
point in D.
(b) Using properties of continuous functions, show that the function f (x) = x1
is continuous on the interval 0 < x < .
(c) Is the converse of (a) always true? That is, every continuous function is
uniformly continuous.

Solution.
(a) Suppose f is uniformly continuous on D. Let a be a point in D. Let  > 0
be given. Then by uniform continuity, there is a > 0 such that for all x, u
in D,

if |x u| < = |f (x) f (u)| < .

73
In particular, the above is true if u = a. Hence, f is continuous at a.
(b) Since g(x) = 1 and h(x) = x are both continuous in the interval 0 < x <
, the ratio function f (x) = x1 is also continuous there.
(c) No. The function f (x) = x1 is continuous in the interval 0 < x < (by
(b)) but not uniformly continuous there (by Exercise 13.2(a))
Exercise 13.4
Show that if f, g : D R are uniformly continuous then f + g : D R is
also uniformly continuous.
Solution.
Let  > 0. Since f is uniformly continuous, there is a 1 > 0 such that

if |x u| < 1 = |f (x) f (u)| < 2
for all x, u in D.
Likewise, since g is uniformly continuous, there is a 2 > 0 such that

if |x u| < 2 = |g(x) g(u)| < 2
for all x, u in D.
Let = min{1 , 2 }. If x, u are in D such that |x u| < then |x u| < 1
and |xu| < 2 . But then |f (x)f (u)| < 2 and |g(x)g(u)| < 2 . Therefore,
|(f + g)(x) (f + g)(u)| =|(f (x) f (u)) + (g(x) g(u))|
|f (x) f (u)| + |g(x) g(u)|
 
< + = .
2 2
It follows that f + g is also uniformly continuous
Exercise 13.5
Let f (x) = x2 . Suppose that there is a > 0 such that |x1 x2 | < for all
real numbers x1 and x2 . In addition, suppose we want |x21 x22 | = 1. That is,
|x1 x2 ||x1 + x2 | = 1. One way to achieve that is by setting x1 x2 = 2 and
x1 + x2 = 2 .
(a) Find x1 and x2 in terms of .
(b) Show that f is not uniformly continuous. Hint: Let  = 12 in Definition
12.
Solution.
(a) Solving the system of equations by the method of elimination we find
x1 = 4 + 1 and x2 = 1 4 .
(b) By (a), letting  = 21 we see that |x1 x2 | = 2 < but |f (x1 ) f (x2 )| =
1 > 21 = . This shows that the given function is not uniformly continuous

74
Exercise 13.6
Give an example of two functions f, g : D R that are uniformly continuous
but the product function f g is not.

Solution.
Let f (x) = g(x) = x. Both functions are uniformly continuous by Exercise
??. However, the product (f g)(x) = x2 is not uniformly continuous by the
previous exercise

Exercise 13.7
Let f, g : D R be uniformly continuous and bounded, say |f (x)| M1
and |g(x)| M2 for all x in D. Let  > 0 be arbitrary.
(a) Show that there is a 1 > 0 such that

if |x u| < 1 = |f (x) f (u)| < 2M2
for all x, u in D.
(b) Show that there is a 2 > 0 such that

if |x u| < 2 = |g(x) g(u)| < 2M1
for all x, u in D.
(c) Show that f g : D R is also uniformly continuous. Note that bound-
edness is crucial in this result. Hint: Note that f (x)g(x) f (u)g(u) =
(f (x) f (u)g(x) + f (u)(g(x) g(u)).

Solution.
(a) Since f is uniformly continuous, there is a 1 > 0 such that

if |x u| < 1 = |f (x) f (u)| < 2M2
for all x, u in D.
(b) Likewise, since g is uniformly continuous, there is a 2 > 0 such that

if |x u| < 2 = |g(x) g(u)| < 2M1
for all x, u in D.
(c)Let = min{1 , 2 }. If x, u are in D such that |x u| < then |x u| < 1
and |x u| < 2 . But then |f (x) f (u)| < 2M 2 and |g(x) g(u)| < 2M 1 .
Therefore,

|f (x)g(x) f (u)g(u)| =|(f (x) f (u))g(x) + f (u)(g(x) g(u))|

|f (x) f (u)||g(x)| + |f (u)||g(x) g(u)|
 
< M2 + M1 = .
2M2 2M1
It follows that f g is also uniformly continuous

75
Exercise 13.8
Suppose that f : D R is uniformly continuous. Let {an } n=1 be a Cauchy
sequence of terms in D.
(a) Let  > 0 be arbitrary. Show that there is a > 0 such that
If |x1 x2 | < = |f (x1 ) f (x2 )| <  for all x1 , x2 in D.
(b) Show that there is a positive integer N such that
If n, m N = |an am | < .
(c) Show that {f (an )}n=1 is a Cauchy sequence in R (and therefore by Ex-
ercise 7.7 is convergent).

Solution.
(a) This is just the definition of uniform continuity.
(b) This is just the definition of Cauchy sequence.
(c) For the given  > 0 we can find a positive integer N such that if n, m N
we have |an am | < by (b). Using (a), |an am | < implies that |f (an )
f (am )| < . Thus, {f (an )}
n=1 is Cauchy in R and therefore is convergent

Exercise 13.9
Consider the function f (x) = tan x on the interval 2 < x < 2 .
(a) Show that the sequence { 2 n1 }
n=1 is convergent.
(b) Show that the sequence in (a) is also

Cauchy.
1
(c) Show that the sequence {f 2 n }n=1 is not Cauchy.
(d) Show that the function f (x) = tan x is not uniformly continuous on the
interval 2 < x < 2 .

Solution.
(a) limn 2 n1 = 2 .

(b) This follows from Exercise 5.2.

1
n1
= limn cot n1 = cot limn n1 =




(c) Since limn f 2 n = limn tan 2
cot 0 which is undefined, the sequence {f
2
n1 }n=1 is not convergent and
therefore is not Cauchy by Exercise 7.7.
(d) This follows from Exercise 13.8

Exercise 13.10
Let f : D R and g : D0 R be two uniformly continuous functions with
the range of f contained in D0 . Looking closely at Exercise 12.3, show that
the composite function g(f (x)) is also uniformly continuous.

76
Solution.
Let  > 0 be given. Since g is uniformly continuous, there is a 0 > 0 such
that if |u v| < 0 then |g(u) g(v)| < , where u and v are in D0 .
Now, since f is uniformly continuous, there is a 00 > 0 such that if |x1 x2 | <
00 then |f (x1 )f (x2 )| < 0 , where x1 and x2 are in D. Let = min{ 0 , 00 }. Let
x1 and x2 be in D such that |x1 x2 | < . Then |x1 x2 | < 00 which implies
that |f (x1 )f (x2 )| < 0 . But f (x1 ) and f (x2 ) are in D0 with |f (x1 )f (x2 )| <
0 . Then |g(f (x1 )) g(f (x2 ))| < . This establishes the fact that g(f (x)) is
uniformly continuous

Exercise 13.11
Consider the function f (x) = sin x defined on the interval 2 < x < 2 .
(a) Use Exercise 9.11(a) to show that | sin x| |x| on the interval 2 < x <

2
.
(b) Using the trigonometric identity sin a sin b = 2 sin ab cos a+b



2 2
and
(a) to show that
| sin a sin b| |a b|.
(c) Show that f is uniformly continuous on the 2 < x < 2 .

Solution.
(a) For 0 x < 2 we know that sin x 0 so that |x| sin x x = |x|.
That is, | sin x| |x|. For 2 < x < 0 we have | sin (x)| | x| which is
the same as | sin x| |x|.
(b) Using (a) and the fact that | cos x| 1 we can write
    
 ab a + b 
| sin a sin b| = 2 sin
 cos
2 2 
     
 a b   a + b 
=2 sin cos
2   2 
  
 a b  |a b|
2 sin 2 = |a b|
2  2

(c)Let  > 0 be given. Choose = . Then for all a and b on the interval
2 < x < 2 we have

if |a b| < = | sin a sin b| |a b| < = 

Thus, f (x) = sin x is uniformly continuous on the interval 2 < x <

2

77
Exercise 13.12
Using Exercise 13.10 and Exercise 13.11, show that the function g(x) = cos x
is uniformly continuous in the interval 2 < x < 2 .

Solution.
First, recall that cos x = sin 2 x . Let f (x) = 2 x and g(x) = sin x.

Then cos x = g(f (x)) is uniformly continuous (by Exercise 13.10) since both
f and g are uniformly continuous( by Exercise 13.11 and Exercise 13.1)

Exercise 13.13
Give an example of two uniformly continuous functions f and g such that
f (x)
g(x)
is not uniformly continuous.

Solution.
The functions f (x) = sin x and g(x) = cos x are uniformly continuous on the
interval 2 < x < 2 (See Exercise 13.11 and Exercise 13.12.) However, the
function fg(x)
(x)
= tan x is not uniformly continuous in the interval 2 < x < 2
(See Exercise 13.9)

Exercise 13.14
Let g : D R be a uniformly continuous function with |g(x)| M > 0 for
1
all x D. Hence, the function g(x) is bounded and g(x) 6= 0 for all x in D.
1
Show that g(x) is uniformly continuous.

Solution.
Let  > 0 be given. By uniform continuity, there is a > 0 such that

if |a b| < = |g(a) g(b)| < M 2 .

Thus, if |a b| < we can also have

   
 1 1   g(b) g(a) 
 g(a) g(b)  =  g(a)g(b) 
   

|g(a) g(b)|
=
|g(a)||g(b)|
|g(a) g(b)| M 2
< =
M2 M2
1
This shows that g(x)
is uniformly continuous

78
Exercise 13.15
Let f, g : D R be two uniformly continuous functions such that f (x) is
bounded and |g(x)| M > 0 for all x D. Show that the function fg(x)
(x)
is
uniformly continuous on D.

Solution.
Note that fg(x)
(x) 1
= f (x) g(x) which is the product of two uniformly continuous
1 f (x)
functions with f (x) and g(x)
bounded. By Exercise 13.7(c), the function g(x)
is uniformly continuous

Exercise 13.16
A function f : D R is said to be Lipschitz if there is a constant K > 0
such that |f (x) f (y)| K|x y| for all x, y D. Show that a Lipschitz
function is uniformly continuous.

Solution.
Let  > 0 be given. Let = K . Then for all x, y D such that |x y| <
we have

|f (x) f (y)| K|x y| < K = .
K
This shows that f is uniformly continuous

79
Solutions to Section 14
Exercise 14.1
(a) Let c0 = a+b 2
. Then either [a, c0 ] or [c0 , b] contains an infinite members of

{vn }n=1 . Lets call the interval [a1 , b1 ]. Show that b1 a1 = ba 2
.
a1 +b1
(b) Let c1 = 2 . Then either [a1 , c1 ] or [c1 , b1 ] contains an infinite members
of {vn } ba
n=1 . Lets call the interval [a2 , b2 ]. Show that b2 a2 = 22 . Compare
a1 and a2 . Compare b1 and b2 .
(c) Let c2 = a2 +b2
2
. Then either [a2 , c2 ] or [c2 , b2 ] contains an infinite members

of {vn }n=1 . Lets call the interval [a3 , b3 ]. Show that b3 a3 = ba 23
. Compare
a1 , a2 and a3 . Compare b1 , b2 and b3 .

Solution.
(a) If a1 = a and b1 = c0 then b1 a1 = a a+b 2
= ba2
. If a1 = c0 and b1 = b
a+b ba
then b1 a1 = b 2 = 2 .
(b) If a2 = a1 and b2 = c1 then b2 a2 = a1 +b
2
1
a1 = b1 a
2
1
= ba
22
. If a2 = c1
a1 +b1 b1 a1 ba
and b2 = b1 then b2 a2 = b1 2 = 2 = 22 . Note that a2 a1 and
b2 b1 .
(c) If a3 = a2 and b3 = c2 then b3 a3 = a2 +b
2
2
a2 = b2 a2
2
= ba
23
. If a3 = c2
a2 +b2 b2 a2 ba
and b3 = b2 then b3 a3 = b2 2 = 2 = 23 . Note that a3 a2 a1
and b3 b2 b1

Exercise 14.2
(a) Show that the sequence {an } n=1 is bounded from above. What is an
upper bound?
(b) Show that there is a constant M such that M = sup{a1 , a2 , }.
(c) Show that a M b.

Solution.
(a) For all n 1, we have an b. That is the sequence {an }n=1 is bounded
from above with an upper bound b.
(b) By the Completeness axiom of R there is a finite number M such that
M = sup{a1 , a2 , }.
(c) Since b is an upper bound of {an }
n=1 and M is the smallest upper bound,
we must have M b. Since M is an upper bound of {an } n=1 we must
have an M for all n 1. Since a an for all n 1 we conclude that
aM b

80
Exercise 14.3
(a) Show that there is > 0 such that for any a x b if |x M | < then
|f (x) f (M )| < 2 .
(b) Show that for all u and v in [a, b] if |u M | < and |v M | < then
|f (u) f (v)| < .

Solution.
(a) This follows from the fact that f is continuous at M.
(b) If u and v are in [a, b] such |u M | < and |v M | < then

|f (u) f (v)| =|(f (u) f (M )) (f (v) f (M ))|

|f (u) f (M )| + |f (v) f (M )|
 
< + =
2 2
Exercise 14.4
(a) Let wn = ba
2n
. Show that limn wn = 0. Hint: Squeeze rule.
(b) Show that there is a positive integer N such that ba
2N
< 2 and |xM | <
2
for all aN x bN .

Solution.
(a) For n 1, we have 2n n. Thus, 0 21n n1 and 0 wn ban
. By the
ba
Squeeze rule we conclude that limn 2n = 0.
(b) From (a), there is a positive integer K such that if n K then ba
2n
< 2 .
Let N be large enough so that N K and |x M | < 2 for all aN x bN

Exercise 14.5
(a) Using Exercise 14.4, show that there is a large n such that n1 < 2 and
aN v n b N .
(b) For the n found in (a), show that |un vn | < n1 < 2 and |vn M | < 2 .
(c) For the n found in (a), Show that |un M | < .
(d) Using (b), (c), and Exericse 14.3(b), show that |f (un ) f (vn )| < .
Conclusion: The result in (d), contradicts (1). Hence, f must be uniformly
continuous.

Solution.
(a) Since [aN , bN ] contains an infinite number of vn , we can choose n large
enough so that n1 < 2 and aN vn bN .
(b) Since aN vn bN , by the previous exercise we obtain |vn M | < 2 < .

81
Moreover, from (??), we have |un vn | < n1 < 2 .
(c) We have |un M | = |(un vn )+(vn M )| |un vn |+|vn M | < 2 + 2 = .
(d) Since |un M | < and |vn M | < 2 < , by Exericse 14.3(b) we obtain
|f (un ) f (vn )| < . But this contradicts (1). Hence, f must be uniformly
continuous

Exercise 14.6
Show that the function f : [0, 1] R defined by f (x) = x is uniformly
continuous.

Solution.
By Exercise 11.15, f is continuous on [0, 1]. Thus, f is uniformly continuous
on [0, 1]

Exercise 14.7
(a) A function f : D R is said to be Lipschitz if there is a constant K > 0
|f (x) f (y)| K|x y| for all x, y D. Show that the function
such that
f (x) = x is not Lipschitz on [0, 1]. Hint: Assume the contrary and get a
contradiction.
(b) Give an example of a uniformly continuous function that is not Lipschitz.
Thus, the converse to Exercise 13.16 is false.

Solution.
(a) Supposethat f is Lipschitz on [0, 1]. Then there is a positive constant K

that | x y| K|x y| for all x, y [0, 1]. Letting y = 0 we obtain
such
x Kx for all x [0, 1]. In particular, x Kx for all 0 < x 1 and
therefore K 1x for all x (0, 1]. Letting x 0+ we see that K , a
contradiction.
(b) Let f : [0, 1] R be defined by f (x) = x. We know from (a) that f is
not Lipschitz. We also know by Exercise 14.6 that f is uniformly continuous

Exercise 14.8
Show, using the definition of uniform continuity (epsilon-delta definition) the
x
function f (x) = x+1 is uniformly continuous on [0, 2].

Solution.
Let x, y [0, 2]. Note that
 
 x y  |x y|
 x + 1 y + 1  = (x + 1)(y + 1) .


82
Let  > 0 be given. Choose < . Since (x + 1)(y + 1) 1 for |x y| < we
have  
 x y  |x y|
 x + 1 y + 1  = (x + 1)(y + 1) |x y| < .


Hence, the given function is uniformly continuous on [0, 2]

Exercise 14.9
Conisder the function f : [0, 1] R defined by
 sin x
x
if 0 < x 1
f (x) =
1 if x = 0

Show that f is uniformly continuous on [0, 1].

Solution.
f (x) is continuous for all x (0, 1]. Since limx0+ f (x) = 1 = f (0) the
function is continuous at x = 0. Thus, the function f is continuous on [0, 1]
and therefore uniformly continuous

Exercise 14.10
Show that the function f : (2, 1] R defined by f (x) = x2 is Lipschitz on
(2, 1].

Solution.
We have

|x2 y 2 | = |x + y||x y| (|x| + |y|)|x y| 4|x y|

83
Solutions to Section 15
Exercise 15.1
Give an example of a continuous f : D R with D a bounded set (i.e.
|x| M for some M > 0 and for all x in D) but f (D) is not bounded.

Solution.
Let D be the open interval (0, 1). Then D is bounded. Define f : D R
by f (x) = x1 . We know from Exercise 13.3(b) that f is continuous on D.
Moreover, f (D) is the open interval (0, ) which is not bounded

Exercise 15.2
Let D be a bounded subset of R with |x| M for all x D. Suppose that
f : D R is uniformly continuous.
(a) Show that there is a > 0 such that if u and v belong to D such that
|u v| < then |f (u) f (v)| < 1.
(b) Let n be a positive integer such that n > 2M
. Divide the interval
[M, M ] into n equal subintervals:[x0 , x1 ], [x1 , x2 ], , [xn1 , xn ]. Show that
xk xk1 < for all k = 1, 2, , n
(c) Let [a1 , b1 ], [a2 , b2 ], , [ak , bk ] be those intervals in (b) that intersect D.
That is, D [a1 , b1 ] [a2 , b2 ] [ak , bk ]. For 1 i k let ui [ai , bi ] D.
Show that if v is in D then there is an 1 i k such that |v ui | < and
|f (v)| < 1 + |f (ui )|.
(d) Show that |f (v)| M for all v in D. That is, f (D) is bounded.

Solution.
(a) Let  = 1. By uniform continuity, there is a > 0 such that if u and v
belong to D such that |u v| < then |f (u) f (v)| < 1
(b) We have xk xk1 = 2M n
< .
(c) Since D [a1 , b1 ] [a2 , b2 ] [ak , bk ], if v is in D then v belongs to
[ai , bi ] for some 1 i k. But then |v ui | < |bi ai | < (by (b)). By (a)
we must have |f (v) f (ui )| < 1. Using the triangle inequality, we conclude
that |f (v)| < 1 + |f (ui )|.
(d) Let M = 1 + |f (u1 )| + |f (u2 )| + + |f (uk )|. By part (c), for any v in D
we have |f (v)| < 1 + |f (ui )| < M. Thus, the range of f is bounded

Exercise 15.3
Show that if f : [a, b] R is continuous then f ([a, b]) is bounded. Hint:
Exercises 14.5 and 15.2.

84
Solution.
Since f is continuous, by Exercise 14.5 it is uniformly continuous. Since [a, b]
is bounded, by Exercise 15.2, f ([a, b]) is bounded

Exercise 15.4
Show that if f : [a, b] R is continuous then inf{f (x) : a x b} and
sup{f (x) : a x b} exist.

Solution.
By Exercise 15.3, f ([a, b]) is bounded. Now, the result follows from the
Completenes Axiom of real numbers

Exercise 15.5
Let f : [a, b] R be continuous. Let I = inf{f (x) : a x b}. Note that I
exists by Exercise 15.4. Suppose that I < f (x) for all x [a, b]. That is, the
infimum can not be attained in [a, b]. Define the function g : [a, b] R by
1
g(x) = .
f (x) I

(a) Show that g is continuous on [a, b].

(b) Show that there is a positive number M such that |g(x)| M for all
x [a, b].
(c) Show that I + M1 is a lower bound of f ([a, b]) and this leads to a contra-
diction.
Conclusion: There must be a number x1 [a, b] such that f (x1 ) = inf{f (x) :
a x b}.

Solution.
(a) Since the constant function w1 (x) = 1 and w2 (x) = f (x) I are con-
tinuous, the function g being the ratio of two continuous functions is also
continuous.
(b) By Exercise 15.3, g([a, b]) is bounded, that is |g(x)| M for all x [a, b].
1
(c) From (b), we have f (x)I M which implies that I + m1 f (x) for all
x [a, b]. But this says that I + M1 is a lower bound of f ([a, b]). Since I is
the largest lower bound, we must have I + M1 I, a contradiction. Hence,
there must be a number x1 [a, b] such that f (x1 ) = inf{f (x) : a x b}

85
Exercise 15.6
Let f : [a, b] R be continuous. Let S = sup{f (x) : a x b}. Note
that S exists by Exercise 15.4. Show that there exists x2 [a, b] such that
f (x2 ) = S. Hint: Mimic Exercise 15.5.

Solution.
Suppose that f (x) < S for all x [a, b]. Define the function g(x) = Sf1 (x) .
Since the constant function w1 (x) = 1 and w2 (x) = S f (x) are continuous,
the function g being the ratio of two continuous functions is also continuous.
By Exercise 15.3, g([a, b]) is bounded, that is |g(x)| M for all x [a, b].
Thus, we have Sf1 (x) M which implies that f (x) S M1 for all x [a, b].
But this says that S M1 is an upper bound of f ([a, b]). Since S is the smallest
upper bound, we must have S < S M1 , a contradiction. Hence, there must
be a number x2 [a, b] such that f (x2 ) = sup{f (x) : a x b}

Exercise 15.7
Let f : [a, b] R be continuous. Let f (a) c f (b).
(a) Let D = {x [a, b] : f (x) c}. Show that D is non-empty and that D
is bounded from above. By the Completeness Axiom of real numbers there
is a number d such that d = sup{x D}.
(b) Show that d [a, b].
(c) Suppose that f (d) > c. Show that there is a > 0 such that if |x d| <
then |f (x) f (d)| < f (d) c.
(d) Show that for x [a, b] and |x d| < we must have f (x) > c. Hint:
Exercise 1.14.
(e) Using (d), show that d is an upper bound of D. Thus, f (d) > c leads
to a contradiction.
(f) Suppose that f (d) < c. Show that there is a > 0 such that if d <
x < d + and x [a, b] we must have f (x) < c.
(g) Show that f (d + 2 ) < c. Why this leads to a contradiction?
Conclusion: We must have f (d) = c.

Solution.
(a) Since a D, D is non-empty. Moreover, since x b for all x D, we see
that D is bounded from above. By the Completeness Axiom of real numbers
there is a number d such that d = sup{x D}.
(b) By the definition of d, we have a x d for all x D. Also, since b
is an upper bound of D and d is the smallest upper bound, we must have

86
d b. Hence, a d b.
(c) Let  = f (d) c > 0. Since f is continuous at d, there is a > 0 such
that if |x d| < then |f (x) f (d)| <  = f (d) c.
(d) If |x d| < then |f (x) f (d)| < f (d) c which is equivalent to
c f (d) < f (x) f (d) or f (x) > c.
(e) If x D and x > d then x d > which implies that f (x) > c a
contradiction. Hence, x d so that d is an upper bound of D. But
then d < d which is impossible. Hence, f (d) > c cannot happen.
(f) If f (d) < c. Letting  = c f (d), we can find > 0 such that if |x d| <
we have |f (x) f (d)| < c f (d) or f (x) < c.
(g) Let x = d + 2 . Then d < x < d + and therefore f (x) < c, that is
d + 2 D. But the definition of d implies that d + 2 < d, a contradiction

Exercise 15.8
Let f : [a, b] R be continuous. By Exercise 15.5, there exist x1 [a, b] and
x2 [a, b] such that m = f (x1 ) = inf{f (x) : x [a, b]} and M = sup{f (x) :
x [a, b]}.
(a) Show that f ([a, b]) [m, M ].
(b) Use Exercise 15.7 (restricted to [x1 , x2 ]) to show that [m, M ] f ([a, b]).
Conclusion: f ([a, b]) = [m, M ].

Solution.
(a) Let y f ([a, b]). Then there is x [a, b] such that y = f (x). But
m f (x) M, that is, m y M. This shows that y [m, M ] and
therefore f ([a, b]) [m, M ].
(b) Let y [m, M ]. By Exercise 15.7 restricted to the interval [x1 , x2 ], there
exists a x1 x x2 b such that f (x) = y. That is, y f ([a, b]). This
shows that [m, M ] f ([a, b])

Exercise 15.9
Prove that there exists a number c 0, 2 such that 2c 1 = sin c2 + 4 .

Solution.
Let f (x) = 2x 1 sin x2 + 4 . Then f (0) = 1 12 < 0 and f 2 =



 2 
1 sin 2 + 4 > 0. By the Intermediate value theorem, there is a
c 0, 2 such that f (c) = 0 or 2c 1 = sin c2 + 4

87
Exercise 15.10
Let f : [a, b] [a, b] be a continuous function. Prove that there is c [a, b]
such that f (c) = c. We call c a fixed point of f. Hint: Intermediate Value
Theorem applied to a specific function F (to be found) defined on [a, b].

Solution.
Define F : [a, b] R by F (x) = x f (x). Then F is continuous on [a, b].
Since a f (a) b and a f (b) b we find F (a) = a f (a) 0 and
F (b) = b f (b) 0. By the Intermediate Value Theorem, there is a c [a, b]
such that F (c) = 0 or c f (c) = 0. Thus, f (c) = c

Exercise 15.11
Using the Intermediate Value Theorem, show that
(a) the equation 3 tan x = 2 + sin x has a solution in the interval (0, 4 ).
(b) the polynomial p(x) = x4 + 2x3 + 2 has at least two real roots.

Solution.
(a) Let f (x) = 3 tan x sin x 2. Then f is continuous on [0, 4 ] and f (0) =
2 < 0, f 4 = 1 12 > 0. By IVT, there is a c (0, 4 ) such that f (c) = 0.

This means, the given equation has at least one solution in the interval (0, 4 ).
(b) Since p(1) = 1 < 0, p(0) = 2 > 0, and p(3) = 25 < 0 there exist at
least two numbers 1 < c1 < 0 < 3 such that f (c1 ) = f (c2 ) = 0

Exercise 15.12
Let f, g : [a, b] R be continuous functions such that f (a) g(a) and
f (b) g(b). Show that there is a c [a, b] such that f (c) = g(c).

Solution.
Let h(x) = f (x) g(x). Then h is continuous on [a, b] with h(a) 0 h(b).
By the IVT, there is a c [a, b] such that h(c) = 0 or f (c) = g(c)

Exercise 15.13
Let f : [a, b] R be continuous such that f (a) a and f (b) b. Prove that
there is a c [a, b] such that f (c) = c. We call c a fixed point of f.

Solution.
Let g(x) = f (x)x. Then g is continuous on [a, b] with g(a) 0 and g(b) 0.
By IVT, there is a c [a, b] such that g(c) = 0. That is, f (c) = c

88
Exercise 15.14
Let f : [a, b] R\Q be continuous. Prove that f must be a constant
function. Hint: Exercise 2.6(c).

Solution.
Suppose for a contradiction that f is not constant. Then, we can find x, y
[a, b] with x < y and such that f (x) 6= f (y). Choose an rational number m
lying between f (x) and f (y). Then, by the intermediate value theorem, there
exists z [x, y] with f (z) = m. Hence, f takes a rational value, contradicting
the hypotheses

Exercise 15.15
Prove that a polynomial of odd degree considered as a function from the reals
to the reals has at least one real root.

Solution.
Let f (x) be a polynomial of odd degree. Then limx f (x) = and
limx f (x) = (or limx f (x) = and limx f (x) = depend-
ing on whether the leading coefficient is positive or negative, respectively).
Hence, there exist a < b in R such that f (a) < 0 and f (b) > 0: Now the
Intermediate Value Theorem applies to give an x (a, b) such that f (x) = 0

Exercise 15.16
Suppose f (x) is continuous on the interval [0, 2] and f (0) = f (2) : Prove
there must be a number c between 0 and 1 so that f (c + 1) = f (c). Hint:
Consider the function g(x) = f (x + 1) f (x) on [0, 1].

Solution.
We let g(x) = f (x + 1) f (x). f (x) is continuous on [0, 1]. Furthermore,

g(0) = f (1) f (0)

and
g(1) = f (2) f (1) = f (0) f (1) = (f (1) f (0)).
If f (1) = f (0) we have obtained the desired conclusion upon taking c = 0.
We therefore assume f (0) 6= f (1). But then g(0) and g(1) have opposite
signs. The Intermediate Value Theorem therefore guarantees the existence
of a number c in the open interval (0, 1) satisfying g(c) = 0. But by definition
of g(x), this means f (c + 1) = f (c)

89
Solutions to Section 16
Exercise 16.1
Consider the function
1


x sin x
if x 6= 0
f (x) =
0 if x = 0

Show that f is not differentiable at a = 0.

Solution.
We have  
0 f (0 + h) f (0) 1
f (0) = lim = lim sin .
h0 h h0 h
By Exercise 8.3, the limit does not exist. Hence, f is not differentiable at 0

Exercise 16.2
Consider the function
1


x2 sin x
if x 6= 0
f (x) =
0 if x = 0

Show that f is differentiable at a = 0. What is f 0 (0)?

Solution.
We have  
0 f (0 + h) f (0) 1
f (0) = lim = lim h sin .
h0 h h0 h
By Exercise 9.4(b), the limit does exist and is equal to 0

Exercise 16.3
Show that f (x) = |x| is not differentiable at 0.

Solution.
We have
f (h) f (0) |h|
f 0 (0) = lim = lim .
h0 h h0 h

By Exercise 8.2, the limit does not exist. Hence f is not differentiable at 0

90
Exercise 16.4
Find the derivative of f
(x) = sin
x. Hint: Recall the trigonometric identity
sin a sin b = 2 cos a+b
2
sin ab
2
and use Exercise 9.11.

Solution.
Using Exercise 9.11 along with the continuity of the cosine function we have
f (x + h) f (x) sin (x + h) sin x
f 0 (x) = lim = lim
h0 h

h0 h
2x+h h
2 cos 2 sin 2
= lim
h0 h
 "
#
sin h2
 
2x + h
= lim cos lim h
= cos x 1 = cos x
h0 2 h0
2

Exercise 16.5
Let f : D R be differentiable at a.
(a) Show that

lim [f (x) f (a)] = lim [f (h + a) f (a)].

xa h0

(b) Show that f is continuous at a. That is,

lim [f (x) f (a)] = 0.

xa

Solution.
(a) Letting h = x a and noting that limxa f (x) = limh0 f (h + a), We
have

lim [f (x) f (a)] = lim [f (h + a) f (a)].

xa h0

(b) We have

lim [f (x) f (a)] = lim [f (h + a) f (a)]

xa h0
f (a + h) f (a)
= lim h
h0
 h 
f (a + h) f (a)
= lim [lim h] = f 0 (a) 0 = 0
h0 h h0

Hence, f is continuous at a

91
Exercise 16.6
Give an example of a function f : D R that is continuous at a but not
differentiable there.

Solution.
The function f (x) = |x| is continuous at 0 but not differentiable there (Ex-
ercise 16.3)

Exercise 16.7
Suppose that f, g : D R are differentiable at a. Show that the functions
f g are also differentiable at a.

Solution.
We have
(f g)(a + h) (f g)(a)
(f g)(a) = lim
h0
 h 
f (a + h) f (a) g(a + h) g(a)
= lim
h0 h h
   
f (a + h) f (a) g(a + h) g(a)
= lim lim
h0 h h0 h
0 0
=f (a) g (a)

Exercise 16.8 (Product Rule)

Suppose that f, g : D R are differentiable at a.
(a) Show that (f g)(a + h) (f g)(a) = [f (a + h) f (a)]g(a + h) + f (a)[g(a +
h) g(a)].
(b) Show that the function f g is also differentiable at a.

Solution.
(a) We have

[f (a + h) f (a)]g(a + h) + f (a)[g(a + h) g(a)] =f (a + h)g(a + h) f (a)g(a + h)

+f (a)g(a + h) f (a)g(a)
=(f g)(a + h) (f g)(a).

92
(b) Using (a) and the continuity of g We have

(f g)(a + h) (f g)(a)
(f g)0 (a) = lim
h0 h
[f (a + h) f (a)]g(a + h) + f (a)[g(a + h) g(a)]
= lim
h0
  h  
f (a + h) f (a) g(a + h) g(a)
= lim [lim g(a + h)] + f (a) lim
h0 h h0 h0 h
0 0
=f (a)g(a) + f (a)g (a)

Exercise 16.9 (Quotient Rule)

Suppose that f, g : D R are differentiable at a with g(a) 6= 0.
(a) Show that
   
f f
g
(a + h) g
(a) f (a + h) f (a) 1 g(a + h) g(a) f (a) 1
= .
h h g(a + h) h g(a) g(a + h)

(b) Show that  0

f f 0 (a)g(a) f (a)g 0 (a)
(a) = .
g g(a)2

Solution.
(a) We have
   
f f
g
(a + h) g
(a) f (a + h)g(a) f (a)g(a + h)
=
h hg(a + h)g(a)
[f (a + h) f (a)]g(a) f (a)[g(a + h) g(a)]
=
hg(a + h)g(a)
f (a + h) f (a) 1
=
h g(a + h)
g(a + h) g(a) f (a) 1

h g(a) g(a + h)

(b) Since g is differentiable at a, it is continuous there. That is, limh0 g(a +

93
h) = g(a). Now,
   
f f
 0
f g
(a + h) g
(a)
(a) = lim
g h0 h
f (a + h) f (a) 1
= lim
h0 h g(a + h)
g(a + h) g(a) f (a) 1
lim
h0 h g(a) g(a + h)
f (a + h) f (a) 1
= lim
h0 h limh0 g(a + h)
f (a) g(a + h) g(a) 1
lim
g(a) h0 h limh0 g(a + h)
0 0
f (a)g(a) f (a)g (a)
=
g(a)2

Exercise 16.10 (Chain Rule)

Let f : D R and g : D0 R be two functions with f (D) D0 . Suppose
that f is differentiable at a and g is differentiable at f (a).
(a) Define w : D0 R by
(
g(y)g(f (a))
yf (a)
if y 6= f (a)
w(y) = 0
g (f (a)) if y = f (a).

Show that w is continuous at f (a). That is,

lim w(f (a) + h) = w(f (a)).

h0

(b) Show that (w f )(x) is continuous at a.

(c) Show that

(g f )(a + h) (g f )(a) f (a + h) f (a)

= (w f )(a + h) .
h h
(d) Show that
(g f )0 (a) = g 0 (f (a)) f 0 (a).

94
Solution.
(a) This follows from

g(h + f (a)) g(f (a))

lim w(h + f (a)) = lim = g 0 (f (a)) = w(f (a))
h0 h0 h
(b) The composition of two conitnuous functions is a continuous function
according to Exercise 12.3(a).
(c) We have

(g f )(a + h) (g f )(a) g(f (a + h)) g(f (a)) f (a + h) f (a)

=
h f (a + h) f (a) h
f (a + h) f (a)
=(w f )(a + h) .
h
(d) We have

g(f (a + h)) g(f (a)) f (a + h) f (a)

(g f )0 (a) = lim
h0 f (a + h) f (a) h
 
f (a + h) f (a)
= lim (w f )(a + h) lim
h0 h0 h
0 0 0
=w(f (a)) f (a) = g (f (a)) f (a)

Exercise 16.11 (Power Rule)

Let f (x) = xn where n is a non-negative integer.
(a) By letting h = ax x, show that

f (ax) f (x)
f 0 (x) = lim .
a1 ax x
(b) What is the quotient of the division of an 1 by a1? Hint: use synthetic
division.
(c) Use (a) and (b) to show that f 0 (x) = nxn1 .

Solution.
(a) From the definition of f 0 (x) we have

f (x + h) f (x)
f 0 (x) = lim .
h0 h

95
Letting h = ax x we see that a 1 as h 0. Thus, with this substitution
we can write
f (ax) f (x)
f 0 (x) = lim .
a1 ax x
(b) Using synthetic division we find

an 1 = (a 1)(1 + a + a2 + + an1 ).

(c) Using (a) and (b) we find

(ax)n xn an 1
f 0 (x) = lim = xn1 lim
a1 ax x a1 a 1
n1
=x lim (1 + a + a + + an1 ) = nxn1
2
a1

Exercise 16.12
(a) Show that the derivative of a constant function is zero and that the
derivative of f (x) = x is f 0 (x) = 1.
(b) Show that the function h(x) = x sin x1 is differentiable for all x 6= 0.

Solution.
(a) Suppose that f (x) = C for all x. Then

f (x + h) f (x)
f 0 (x) = lim
h0 h
C C
= lim =0
h0 h
Now, suppose that f (x) = x. Then

f (x + h) f (x)
f 0 (x) = lim
h0 h
x+hx
= lim =1
h0 h
(b) Let f (x) = x1 . Then f is differentiable for all x 6= 0. Let g(x) = sin x.
Then g(x) is differentiable for all x. But h(x) = g(f (x)) so that by the chain
rule h is differentiable for all x 6= 0

Exercise 16.13

Let f (x) = 2x 1. Find f 0 (2) by using only the definition of derivative.

96
Solution.
We have
f (2 + h) f (2)
f 0 (2) = lim
h0
h
2h + 3 3
= lim
h0
h
( 2h + 3 3)( 2h + 3 + 3)
= lim
h0 h( 2h + 3 + 3)
2h
= lim
h0 h( 2h + 3 + 3)

2 3
= lim =
h0 2h + 3 + 3 3

Exercise 16.14
Let 
2x + 5 if x 1
f (x) =
9x2 2 if x > 1.
Show that f (x) is continuous but not differentiable at x = 1.

Solution.
Since
lim f (x) = lim (2x + 5) = 7
x1 x1

and
lim f (x) = lim+ (9x2 2) = 7
x1+ x1

we conclude that f is continuous at x = 1.

Now, since

f (1 + h) f (1) 2(1 + h) + 5 7
lim = lim =2
h0 h h0 h
and
f (1 + h) f (1) 9(1 + h)2 2 7
lim+ = lim+ = lim+ (18 + 9h) = 18
h0 h h0 h h0

the function is NOT differentiable at x = 1

97
Exercise 16.15
Find constants a and b such that the piecewise defined function
 2
ax 4 if x 1
f (x) =
bx + a if x > 1
is differentiable at x = 1.

Solution.
Since f is differentiable at x = 1, it is continuous there. Thus, limx1 f (x) =
limx1+ f (x). That is, a 4 = b + a and this implies that b = 4. Now, since
f is differentiable at 1 we must have
f (1 + h) f (1) f (1 + h) f (1)
lim = lim+ .
h0 h h0 h
That is, 4 = 2a or a = 2

Exercise 16.16
Let f (x) = x2 cos x1 if x 6= 0 and f (0) = 0. Show that f is differentiable at

x = 0 and find f 0 (0).

Solution.
We have  
0 f (h) f (0) 1
f (0) = lim = lim h cos = 0.
h0 h h0 h
Note that the cosine function is bounded and limh0 h = 0

Exercise 16.17
(a) Let f (x) = xn with n a negative integer. Prove that f 0 (x) = nxn1 .
p
(b) Let f (x) = x q where p and q are integers with q 6= 0. Prove that f 0 (x) =
p p
p q 1
q
x . Hint: Let y = x q so that y q = xp and use Exercise 16.10.

Solution.
(a) We have f (x) = 1
xn
so that by the quotient rule we obtain f 0 (x) =
n1
(n)x
(xn )2
= nxn1 .
p
(b) Let y = x q . The y q = xp . Differentiate both sides and use (a) and Exercise
16.10 we find qy q1 y 0 = pxp1 . Thus,
p p p p p
f 0 (x) = y 0 = xp1 y 1q = xp1 x q p = x q 1
q q q

98
Exercise 16.18
We define the number e to be the unique number satisfying

eh 1
lim = 1.
h0 h
It is an irrational number whose value is approximately 2.718281828459045.
Define the function f (x) = ex . Find f 0 (x) using the definition of the deriva-
tive.

Solution.
We have
ex+h ex eh 1
f 0 (x) = lim = ex lim = ex
h0 h h0 h
Exercise 16.19
The natural logarithmic function is the function f (x) = ln x defined
as follows: y = ln x if and only if x = ey . Find the derivative of f. Hint:
Differentiate x = ey with respect to x.

Solution.
Differentiating x = ey with respect to x we obtain 1 = y 0 ey and therefore
y 0 = e1y = x1

Exercise 16.20
Consider the function f (x) = xn where n is a real number.
(a) Suppose that x > 0 and x in the domain of f. Using the fact that
xn = en ln x , show that f 0 (x) = nxn1 .
(b) Suppose that x < 0 and x in the domain of f. Show that f 0 (x) = nxn1 .
Hint: xn = (1)n (x)n .

Solution.
(a) We have f 0 (x) = eln x (n ln x)0 = nx xn = nxn1 .
(b) We have f 0 (x) = (1)n n(x)n1 (1) = (1)n+1 n(x)n1 = (1)n1 n(x)n1 =
nxn1

99
Solutions to Section 17
Exercise 17.1
(a) Find the local extrema (if they exist) of the function f (x) = |x|.
(b) Find the local extrema (if they exist) of the function f (x) = x3 .
(c) Find the local extrema (if they exist) of the function f (x) = x on the
interval [0, 1].

Solution.
(a) Since |x| 0 for all x, we find that f has a local minimum at x = 0.
However, f has no local maximum since f is always increasing for x 0 and
always decreasing for x < 0.
(b) The graph of f (x) = x3 is always increasing so that f has no local
extrema.
(c) The graph is a straight line that is rising to the right. Thus, f (x) has a
local minimum at x = 0 and a local maximum at x = 1

Exercise 17.2
Let f : [a, b] R. Suppose that c (a, b) is a local maximum (or local
minimum) of f such that f 0 (c) exists. Let  > 0 such that f (x) f (c) for
all x (c , c + ) [a, b].
(a) Let h > 0 be small enough so that c + h (c , c + ). Using Exercise
9.8, show that f 0 (c) 0.
(b) Leth < 0 be large enough so that c + h (c , c + ). Using Exercise
9.8, show that 0 f 0 (c) and therefore f 0 (c) = 0.

Solution.
f (c+h)f (c)
(a) Since c + h (c , c + ), we have f (c + h) f (c). Thus, h
0.
By Exercise 9.8, we have
f (c + h) f (c)
f 0 (c) = lim+ lim+ 0 = 0.
h0 h h0

f (c+h)f (c)
(b) Since c + h (c , c + ), we have f (c + h) f (c). Thus, h
0.
By Exercise 9.8, we have
f (c + h) f (c)
f 0 (c) = lim lim 0 = 0.
h0 h h0

From (a) and (b) we conclude that f 0 (c) = 0

100
Exercise 17.3
The condition a < c < b is critical in the previous exercise. Give an example
of a function f : [a, b] R such that either a or b is a local extremum but
with non-zero derivative there.

Solution.
See Exercise 17.1(c). f has a local minimum at 0 with f 0 (0) = 1 6= 0 and a
local maximum at 1 with f 0 (1) = 1 6= 0

Exercise 17.4
Suppose f : [a, b] R is continuous. Then there exists x1 , x2 [a, b] such
that
f (x1 ) f (x) f (x2 )
for all x [a, b]. Show that x1 and x2 are either the endpoints of [a, b] or
critical points of f in a < x < b.

Solution.
Since x1 [a, b], either x1 = a, x1 = b, or a < x1 < b. If a < x1 < b then by
Exercise 17.2, f 0 (x1 ) = 0. That is, x1 is a critical point of f. Similar argument
holds for x2

Exercise 17.5 (Rolles Theorem)

Suppose f : [a, b] R is continuous for a x b and differentiable for
a < x < b. By Exercise 15.8 there exist x1 , x2 [a, b] such that f (x1 )
f (x) f (x2 ) for all x [a, b]. Suppose that f (a) = f (b).
(a) Show that if f (x) = C for all a x b then there is at least a number
a < c < b such that f 0 (c) = 0.
(b) Suppose that f is a non-constant function. Let d [a, b] such that
f (d) 6= f (a). Show that if f (d) < f (a) then a < x1 < b. What can you say
about the value of f 0 (x1 ).
(c) Show that if f (a) < f (d) then a < x2 < b. What can you say about the
value of f 0 (x2 ).

Solution.
(a) If f (x) = C for all a x b then f 0 (x) = 0 for all a x b. In
particular, f 0 a+b = 0 with a < a+b < b. So c = a+b


2 2 2
.
(b) Suppose f (d) < f (a). If x1 = a then f (d) < f (a) = f (x1 ) which is
impossible since f (x1 ) f (x) for all x [a, b]. If x1 = b then f (d) < f (a) =

101
f (b) = f (x1 ) which is again impossible. So we must have a < x1 < b. Now,
by Exercise 17.2 we must have f 0 (x1 ) = 0.
(c) Suppose f (a) < f (d). If x2 = a then f (x2 ) < f (a) which is impossible
since f (x) f (x2 ) for all x [a, b]. If x2 = b then f (x2 ) = f (b) < f (d) which
is again impossible. So we must have a < x2 < b. Now, by Exercise 17.2 we
must have f 0 (x2 ) = 0

Exercise 17.6
Find the number
c of Rolles theorem for the function f : [0, 1] R defined
by f (x) = x x.

Solution.
The function f satisfies the conditions of Rolles theorem. We have 0 =
f 0 (c) = 21 c 1. Solving this equation for c we find c = 41

Exercise 17.7
Assume a0 , a1 , , an are real numbers such that
an an1 a1
+ + + + a0 = 0
n+1 n 2
Show that the polynomial function

f (x) = an xn + an1 xn1 + + a1 x + a0

has at least one root in (0, 1).

Solution.
Let
an n+1 an1 n a1
F (x) = x + x + + x2 + a0 x.
n+1 n 2
Note that F is continuous in [0, 1] and differentiable in (0, 1) with derivative
F 0 (x) = f (x). Moreover, F (0) = F (1) = 0. By Rolles theorem, there is a
c (0, 1) such that F 0 (c) = 0. Hence, f (c) = 0

Exercise 17.8
(a) Show that the function f (x) = x3 4x2 3x + 1 has a root in [0, 2].
(b) Use Rolles theorem to show that there is exactly one root in [0, 2].

102
Solution.
(a) We have f (0) = 1 > 0 and f (2) = 13 < 0 so that by IVT there is a
root in [0, 2].
(b) Suppose that x1 and x2 are two roots of f in [0, 2]. Then by Rolles
theorem we must have c (0, 2) such that f 0 (c) = 0. But the solutions to
f 0 (x) = 0 = 3x2 8x 3 are x = 3 and x = 13 where neither is in [0, 2].
Hence, f has exactly one solution in [0, 2]

Exercise 17.9
Let f, g : R R be differentiable, and let a, b R be such that a < b. Show
that there is a c (a, b) such that

f 0 (c)[g(b) g(a)] = g 0 (c)[f (b) f (a)].

Hint: Apply Rolles theorem to the function h(x) = f (x)[g(b) g(a)]

g(x)[f (b) f (a)].

Solution.
Let h; [a, b] R be the function defined by

h(x) = [f (b) f (a)]g(x) [g(b) g(a)]f (x).

Then h is continuous on [a, b] and differentiable in a < x < b with derivative

h0 (x) = [f (b) f (a)]g 0 (x) [g(b) g(a)]f 0 (x).

Moreover
h(a) = h(b) = f (b)g(a) g(b)f (a).
By Rolles theorem there is a a < c < b such that h0 (c) = 0. That is

[f (b) f (a)]g 0 (c) [g(b) g(a)]f 0 (c) = 0

or
[f (b) f (a)]g 0 (c) = [g(b) g(a)]f 0 (c)

Exercise 17.10
Suppose f : [a, b] R is continuous for a x b and differentiable for
a < x < b. Show that there is a < c < b such that
f (b) f (a)
f 0 (c) = .
ba

103
Hint: Apply Rolles theorem to the function g : [a, b] R defined by
 
f (b) f (a)
g(x) = f (x) f (a) (x a).
ba

Solution.
The function g(x) is continuous on [a, b] being a combination of continuous
functions on [a, b]. Furthermore, g(x) is differentiable for a < x < b with
derivative
f (b) f (a)
g 0 (x) = f 0 (x) .
ba
Also, g(a) = g(b) = 0. By Exercise 17.5, there is a < c < b such that g 0 (c) = 0
which is equivalent to
f (b) f (a)
f 0 (c) =
ba

104
Solutions to Section 18
Exercise 18.1 (Mean Value Theorem)
Suppose f : [a, b] R is continuous for a x b and differentiable for
a < x < b. Show that there is a < c < b such that
f (b) f (a)
f 0 (c) = .
ba
Hint: Use Exercise 17.5 with the function g : [a, b] R defined by
 
f (b) f (a)
g(x) = f (x) f (a) (x a).
ba

Solution.
The function g(x) is continuous on [a, b] being a combination of continuous
functions on [a, b]. Furthermore, g(x) is differentiable for a < x < b with
derivative
f (b) f (a)
g 0 (x) = f 0 (x) .
ba
Also, g(a) = g(b) = 0. By Exercise 17.5, there is a < c < b such that g 0 (c) = 0
which is equivalent to
f (b) f (a)
f 0 (c) =
ba
Exercise 18.2 (Cauchy Mean Value Theorem)
Suppose f, g : [a, b] R are continuous for a x b and differentiable for
a < x < b. Show that there is a < c < b such that

[g(b) g(a)]f 0 (c) = [f (b) f (a)]g 0 (c).

Hint: Use Exercise 17.5 with the function h : [a, b] R defined by

h(x) = [f (b) f (a)]g(x) [g(b) g(a)]f (x).

Solution.
Let h; [a, b] R be the function defined by

h(x) = [f (b) f (a)]g(x) [g(b) g(a)]f (x).

105
Then h is continuous on [a, b] and differentiable in a < x < b with derivative

h0 (x) = [f (b) f (a)]g 0 (x) [g(b) g(a)]f 0 (x).

Moreover
h(a) = h(b) = f (b)g(a) g(b)f (a).
By Rolles theorem there is a a < c < b such that h0 (c) = 0. That is

[f (b) f (a)]g 0 (c) [g(b) g(a)]f 0 (c) = 0

or
[f (b) f (a)]g 0 (c) = [g(b) g(a)]f 0 (c)

Exercise 18.3
Let f : [a, b] R be continuous for a x b and differentiable for a <
x < b. We say that f is one-to-one if and only if for any a x1 b and
a x2 b such that f (x1 ) = f (x2 ) we must have x1 = x2 . Suppose that
f 0 (x) 6= 0 for all a < x < b.
(a) Let a x1 b and a x2 b such that f (x1 ) = f (x2 ). Show that
if x1 < x2 then there is a < x1 < c < x2 < b such that f 0 (c) = 0 which
contradicts the assumption that f 0 (x) 6= 0 for all a < x < b. Hint: Use the
Mean Value Theorem on the interval [x1 , x2 ]..
(b) Answer the same question for x2 < x1 .
Conclusion: We must have x1 = x2 . This shows that f is 1-1.

Solution.
(a) Applying the MVT on the interval x1 x x2 , we can find a number
x1 < c < x2 such that f (x2 )f (x1 ) = f 0 (c)(x2 x1 ). Since f (x1 ) = f (x2 ) = 0
we obtain (x2 x1 )f 0 (c) = 0. Since x1 6= x2 we find that f 0 (c) = 0 which
contradicts the assumption on f 0
(b) Argument similar to (a)

Exercise 18.4
Let f : [a, b] R be continuous for a x b and differentiable for a < x <
b. We say that f is increasing in [a, b] if and only if for every x1 and x2 in
[a, b], if x1 x2 then f (x1 ) f (x2 ). Show that if f 0 (x) 0 for all a < x < b
then f (x) is increasing in [a, b]. Hint: Use the MVT restricted to the interval
[x1 , x2 ].

106
Solution.
Let x1 , x2 [a, b]. Clearly, if x1 = x2 then f (x1 ) = f (x2 ). So assume that
x1 < x2 . By the MVT there is a x1 < c < x2 such that f (x2 ) f (x1 ) =
f 0 (c)(x2 x1 ) 0 which implies that f (x1 ) f (x2 ). Thus, we have shown
that x1 x2 implies f (x1 ) f (x2 ). That is, f is increasing in [a, b]

Exercise 18.5
Consider Case (i). We have either f (y) < f (x) < f (z) or f (z) < f (x) < f (y).
(a) Suppose that f (z) < f (x) < f (y). Use the Intermediate Value theorem
restricted to [y, z] to show that such a double inequality can not occur.
(b) Suppose that f (x) < f (z) < f (y). Use the Intermediate Value theorem
restricted to [x, y] to show that such a double inequality can not occur.
We conclude that Case (i) does not hold.

Solution.
(a) If f (z) < f (x) < f (y), we can apply the Intermediate value theorem to
[y, z] to find y < w < z such that f (w) = f (x). Since f is one-to-one we
must have w = x < y which contradicts the inequality y < w.
(b) If f (x) < f (z) < f (y), we can apply the Intermediate value theorem to
[x, y] to find x < w < y such that f (w) = f (z). Since f is one-to-one we
must have w = z < y which contradicts the inequality y < z

Exercise 18.6
Consider Case (ii). We have either f (y) < f (x) < f (z) or f (y) < f (z) <
f (x).
(a) Suppose that f (y) < f (x) < f (z). Use the Intermediate Value theorem
restricted to [y, z] to show that such a double inequality can not occur.
(b) Suppose that f (y) < f (z) < f (x). Use the Intermediate Value theorem
restricted to [x, y] to show that such a double inequality can not occur.
We conclude that Case (ii) does not hold.

Solution.
(a) If f (y) < f (x) < f (z), we can apply the Intermediate value theorem to
[y, z] to find y < w < z such that f (w) = f (x). Since f is one-to-one we
must have w = x < y which contradicts the inequality y < w.
(a) If f (y) < f (z) < f (x), we can apply the Intermediate value theorem to
[x, y] to find x < w < y such that f (w) = f (z). Since f is one-to-one we
must have w = z < y which contradicts the inequality y < z

107
Exercise 18.7
Suppose that f : [a, b] R is differentiable such that f 0 (x) 6= 0 for all
a < x < b. We know from the above discussion that f is monotone.
(a) Show that if f is increasing in [a, b] then f 0 (x) 0 for all a x b. Hint:
Let x [a, b) and choose h > 0 small enough so that x + h [a, b). If x = b,
choose h < 0 so that b + h < b. Now use the definition of the derivative.
(b) Show that if f is decreasing in [a, b] then f 0 (x) 0 for all a x b.

Solution.
(a) Let x [a, b). Choose h > 0 so that x + h [a, b). Since x < x + h and
f is increasing, we find that f (x+h)f
h
(x)
0. Thus,

f (x + h) f (x)
f 0 (x) = lim 0.
h0 h
If x = b choose h < 0 so that b + h < b. In this case, f (b + h) f (b) and
f (b + h) f (b)
f 0 (b) = lim 0.
h0 h
(b) Let x [a, b). Choose h > 0 so that x + h [a, b). Since x < x + h and
f is decreasing, we find that f (x+h)f
h
(x)
0. Thus,

f (x + h) f (x)
f 0 (x) = lim 0
h0 h
If x = b choose h < 0 so that b + h < b. In this case, f (b + h) f (b) and
f (b + h) f (b)
f 0 (b) = lim 0
h0 h
Exercise 18.8
Let f : [a, b] R be continuous for a x b and differentiable for a <
x < b. We say that f is a constant function on [a, b] if and only if there is a
constant C such that f (x) = C for all a x b. Suppose that f 0 (x) = 0 for
all a < x < b.
Let x1 and x2 be any two numbers in the interval [a, b] with x1 < x2 . Suppose
that f (x1 ) 6= f (x2 ). Show that by applying the Mean Value Theorem on the
interval [x1 , x2 ] we obtain the contradiction f (x1 ) = f (x2 ). Thus, we must
have f (x1 ) = f (x2 ) = C for any x1 and x2 in [a, b]. That is, f (x) = C for all
a x b.

108
Solution.
Applying the MVT on the interval x1 x x2 , we can find a number
x1 < c < x2 such that f (x2 ) f (x1 ) = f 0 (c)(x2 x1 ). Since f 0 (c) = 0 we
obtain f (x1 ) = f (x2 ), a contradiction. Since x1 and x2 were arbitrary, we
have f (x) = C for all x [a, b]

Exercise 18.9
Let f : [a, b] R be continuous for a x b and differentiable for a < x <
b. Suppose that f 0 (x) = g 0 (x) for all a < x < b. Show that f (x) = g(x) + C
for all a x b, where C is a constant. Hint: Exercise 18.8

Solution.
Let h(x) = f (x) g(x). Then h(x) is continuous in [a, b] being the difference
of two continuous functions and h0 (x) = 0 for all a < x < b. By Exercise
18.8, there is C such that h(x) = C for all a x b or equivalently
f (x) = g(x) + C for all a x b

Exercise 18.10
Let f : [a, b] R be continuous for a x b and differentiable for a < x <
b. We say that f is decreasing in [a, b] if and only if for every x1 and x2 in
[a, b], if x1 x2 then f (x1 ) f (x2 ). Show that if f 0 (x) 0 for all a < x < b
then f (x) is decreasing in [a, b]. Hint: Use the MVT restricted to the interval
[x1 , x2 ].

Solution.
Let x1 , x2 [a, b]. Clearly, if x1 = x2 then f (x1 ) = f (x2 ). So assume that
x1 < x2 . By the MVT there is a x1 < c < x2 such that f (x2 ) f (x1 ) =
f 0 (c)(x2 x1 ) 0 which implies that f (x1 ) f (x2 ). Thus, we have shown
that x1 x2 implies f (x1 ) f (x2 ). That is, f is decreasing in [a, b]

Exercise 18.11
Consider the function f (x) = (1 + x)p where 0 < p < 1. Let h > 0.
(a) Apply the MVT to the interval [0, h] to show that f (h) = p(1 + t)p1 h + 1
for some 0 < t < h.
(b) Use (a) to show that (1 + h)p < 1 + ph.
In annuity theory, (1 + h)p may represent compound interest and 1 + ph
represent simple interest. A result in annuity theory says that for time p
less than a year compound interest formula can be estimated by the simple
interest formula.

109
Solution.
(a) Applying the mean value theorem to the interval [0, h], we can find a
0 < t < h such that f (h) f (0) = f 0 (t)h or f (h) 1 = p(1 + t)p1 h. Hence,
f (h) = (1 + h)p = p(1 + t)p1 h + 1.
(b) Since t > 0, we have 1 + t > 1 (1 + t)1p > 1 (1 + t)p1 < 1
p(1 + t)p1 h < ph 1 + p(1 + t)p1 h < 1 + ph. Hence, (1 + h)p < 1 + ph

Exercise 18.12
Suppose that f : [a, b] R is differentiable in [a, b]. Let be a real number
such that either f 0 (a) < < f 0 (b) or f 0 (b) < < f 0 (a).
(a) Define g(x) = f (x) x. Show that if f 0 (a) < < f 0 (b) then g 0 (x)
changes sign between a and b.
(b) Establish the same result for f 0 (b) < < f 0 (a).
(c) Show that the condition g 0 (c) 6= 0 for all c [a, b] leads to a contradiction.
Hint: Exercise 18.7. Conclude that there must be a a < c < b such that
f 0 (c) = .

Solution.
(a) Note that g is continuous in [a, b] and differentiable there with derivative
g 0 (x) = f 0 (x) . Since f 0 (a) < < f 0 (b), we find g 0 (a) = f 0 (a) < 0 <
g 0 (b) = f 0 (b) . So g 0 changes sign from negative to positive.
(b) Since f 0 (b) < < f 0 (a), we find g 0 (b) = f 0 (b) < 0 < g 0 (a) = f 0 (a) .
So g 0 changes sign from positive to negative.
(c) If g 0 (c) 6= 0 for all c [a, b] then by Exercise 18.7 either g 0 is always
nonnegative in [a, b] or always nonpositive which contradict (a) and (b). We
conclude that there must be a a < c < b such that g 0 (c) = 0 which is the
same as f 0 (c) =

Exercise 18.13
Let f, g : [a, b] R be two differentiable functions on [a, b] such that f (a) =
g(a). Show that if f 0 (x) = g 0 (x) for all x (a, b) then f (x) = f (x) for all
x [a, b]. Hint: Exercise 18.8.

Solution.
Let F : [a, b] R be given by F (x) = f (x) g(x). Then F 0 is differentiable
on [a, b] and F 0 (x) = 0 for all x (a, b). By Exercise 18.8, there is a constant
C such that F (x) = C for all x [a, b]. But F (a) = 0 so that C = 0. Thus,
F (x) = 0 for all x [a, b]. This is equivalent to f (x) = g(x) for all x [a, b]

110
Exercise 18.14
Let f : R R be differentiable such that |f 0 (x)| < 1 for all x R. Show
that f can have at most one fixed point. That is, There is at most one c R
such that f (c) = c. Hint: Mean Value Theorem.

Solution.
Suppose the contrary. Let a, b R such that a < b, f (a) = a, and f (b) = b.
We have that f is continuous in [a, b] and differentiable in (a, b). By the
MVT, there is a c (a, b) such that

f (b) f (a) ba
f 0 (c) = = = 1.
ba ba
This is impossible since |f 0 (x)| < 1 for all x R. We conclude that f has at
most one fixed point

Exercise 18.15
Let f : R R be differentiable everywhere and that f 0 (a) < 0 and f 0 (b) > 0
for some a < b. Prove that there is a c (a, b) such that f 0 (c) = 0.

Solution.
This is just Exercise 18.12 with = 0

Exercise 18.16
Let f : R R be differentiable and |f 0 (x)| K < 1 for all x R. Let
a0 R. Define the numbers an = f (an1 ).
(a) Show that |an+1 an | K n |a1 a0 | for all n N.
(b) Show that for all m, n N such that m > n we have

Kn
|am an | .
1K
Solution.
(a) By the MVT there is a c1 (a1 , a0 ) such that f (a1 ) f (a0 ) = f 0 (c1 )(a1
a0 ). Thus, |a2 a1 | K|a1 a0 | since |f 0 (c1 )| K. Likewise, we can write
|a3 a2 | K|a2 a1 | K 2 |a1 a0 |. Now, suppose that |an an1 |
K n |a1 a0 |. Then |an+1 an | K|an an1 | K n+1 |a1 a0 |.
(b) Let m, n N such that mP > n. Then we have |am an | |am am1 | +
Kn
+ |am am1 | |a1 a0 | m i
i=n K = 1K |a1 a0 |

111
Exercise 18.17
a b b


Show that if 0 < a < b then 1 b
< ln a
< a
1. Hint: Apply the MVT
for the function f (x) = ln x.

Solution.
The function f (x) = ln x is continuous on [a, b] and differentiable in (a, b).
By the Mean value theorem there is a a < c < b such that f 0 (c) = ln bln
ba
a
.
Thus,
1 1 ln b ln a 1
< = <
b c ba a
or  
a b b
1 < ln < 1
b a a

112
Solutions to Section 19
Exercise 19.1
Let f, g : [a, b] R be continuous on [a, b] and differentiable in a < x < b
with g 0 (x) 6= 0 for all a < x < b. Suppose that f (c) = g(c) = 0 for some
a c b. Also, suppose that

f 0 (x)
lim = A.
xc g 0 (x)

(a) Let {cn }

n=1 [a, b] be an arbitrary sequence with the properties cn 6= c
for all n 1 and limn cn = c. Show that there is a dn between cn and c
such that
[f (cn ) f (c)]g 0 (dn ) = [g(cn ) g(c)]f 0 (dn ).
(b) Show that dn 6= c for all n 1 and limn dn = c.
(c) Show that g(dn ) 6= g(c) for all n 1. Hint: Exercise 18.3.
(d) Show that
f 0 (dn ) f (cn )
0
= .
g (dn ) g(cn )
f 0 (dn )
(e) Show that limn g 0 (dn )
= A. Hint: See Exercise 10.1.
f (cn )
(f) Show that limn g(cn ) = A.
(g) Show that limxc fg(x)
(x)
= A.

Solution.
(a) Since both f and g are continuous in the closed interval with endpoints
cn and c and differentiable in the open interval with the same endpoints, we
can use Exercise 13.5 to find a point dn in that interval such that

[f (cn ) f (c)]g 0 (dn ) = [g(cn ) g(c)]f 0 (dn ).

(b) Since dn is in the open interval with endpoints cn and c we have dn 6= c

for all n 1. Also, by the Squeeze rule we have limn dn = c.
(c) Since g 0 (x) 6= 0 for all a < x < b, by Exercise 18.3 we find that g is
one-to-one. Since dn 6= c for all n 1, we must have g(dn ) 6= g(c) for all
n 1.
(d) This follows from (a) and the assumption that f (c) = g(c) = 0.

113
 f 0 (x)
(e) Since limxc g 0 (x)
= A, dn 6= c, dn c, we can apply Exercise 10.1 to
write
f 0 (dn )
lim = A.
n g 0 (dn )

(f) Using (d) and (e) the result follows.

(g) The result follows from Exercise 10.1

Exercise 19.2
Find
x 2+ x2
lim .
x2 x2 4
Solution.
Let f (x) = x 2 + x 2 and g(x) = x2 4. Both functions are
continuous in [2, 3] and differentiable in 2 < x < 3. Moreover, f (2) = g(2) =
0. Applying LHopitals rule we find
1 1
x 2+ x2
2 x
+ 2x2
lim = lim
x2 x2 4 x2 x
x2 4

( x 2 + x)( x2 4)
= lim
x2 2x x x 2

( x 2 + x)( x + 2)
= lim
x2 2x x

2 2 1
= =
4 2 2

Exercise 19.3
Let f : [a, b] R be a one-to-one function.
(a) Define g : f ([a, b]) [a, b] by g(y) = x if and only if f (x) = y. Show that
g is indeed a function. That is, if y1 , y2 f ([a, b]) are such that y1 = y2 then
g(y1 ) = g(y2 ).
(b) Show that f (g(y)) = y for all y f ([a, b]) and g(f (x)) = x for all
x [a, b]. Thus, conclude that f is invertible.

Solution.
(a) Define g : f ([a, b]) [a, b] by g(y) = x if and only if f (x) = y. Then g
is a well-defined function: Suppose that y1 , y2 f ([a, b]) such that y1 = y2 .

114
Let x1 , x2 [a, b] such that f (x1 ) = y1 and f (x2 ) = y2 . Then x1 = g(y1 )
and x2 = g(y2 ). Moreover, we have f (x1 ) = f (x2 ). Since f is one-to-one, we
conclude that x1 = x2 which implies that g(y1 ) = g(y2 ).
(b) Let y f ([a, b]). Then f (x) = y for some a x b. By the definition
of g we have g(y) = x. Thus, f (g(y)) = f (x) = y. Likewise, let x [a, b].
Then y = f (x) f ([a, b]). Hence, g(f (x)) = g(y) = x. We conclude that f 1
exists

Exercise 19.4
Let f : [a, b] R be continuous in [a, b] and differentiable in [a, b] with
f 0 (x) 6= 0 for all a < x < b. Let the range of f be denoted by [m, M ].
(a) Show that f is one-to-one, monotone, and invertible with inverse f 1 :
[m, M ] [a, b].
(b) Assume that f is strictly increasing. That is, if x1 < x2 then f (x1 ) <
f (x2 ). In this case, [m, M ] = [f (a), f (b)]. Let f (a) < y0 < f (b). Show that
there is a a < x0 < b such that f (x0 ) = y0 .
(c) Let  > 0 be given. Let 1 = min{, x0 a, b x0 }. Show that if x satisfies
|x x0 | < 1 then a < x < b and |x x0 | < .
(d) Let y1 = f (x0 1 ) and y2 = f (x0 + 1 ). Show that f [(x0 1 , x0 + 1 )] =
(y1 , y2 ).
(e) Choose a > 0 so that (y0 , y0 + ) (y1 , y2 ). Show that if |y y0 | <
then |f 1 (y)f 1 (y0 )| < . This shows that f 1 is continuous in (f (a), f (b)).
(f) Show that f 1 is right continuous at f (a) and left continuous at f (b).
We conclude from this problem that f 1 is continuous on the closed interval
[f (a), f (b)].

Solution.
(a) By Exercise 18.3, f is one-to-one in [a, b]. By Exercises 18.5 and 18.6, f is
monotone. Now, by Exercise 19.3, f is invertible with inverse f 1 : [m, M ]
[a, b].
(b) Since f (a) < y0 < f (b), by the IVT, there is a x0 [a, b] such that
f (x0 ) = y0 . Since f (a) < f (x0 ) < f (b), we must have a < x0 < b.
(c) Suppose that |x x0 | < 1 . Since 1  we can write that |x x0 | < .
Since |x x0 | is less than or equal to both x0 a and b x0 , we conclude
that a < x < b.
(d) First we show that x0 1 [a, b]. Indeed, we have a a + 1 x0 <
b b + 1 . This show that x0 1 [a, b]. Now, let y f [(x0 1 , x0 + 1 )].
Then y = f (x) for some x (x0 1 , x0 + 1 ). Since f is strictly increasing

115
we find y1 < y < y2 . That is, y (y1 , y2 ). The converse is similar.
(e) Suppose that |y y0 | < . Then y (y1 , y2 ) which implies that f 1 (y)
(x0 1 , x0 + 1 ). Hence, |f 1 (y) f 1 (y0 )| < 1 . Thus, we have shown
that for any positive number , we can find a number > 0 such that if
|y y0 | < then |f 1 (y) f 1 (y0 )| < . This says that f 1 is continuous in
(f (a), f (b)).
(f) Let  > 0 be given. There is a 1 > 0 such that if x a < 1 then
f (x) f (a) < . Let 2 = min 1 , . Define = f (a + 2 ) f (a). Suppose
that y f (a) < . Then there is a < x < a + 2 such that f (x) = y. Thus,
|f 1 (y) a| = |x a| < 2 . This shows that f 1 is continuous at f (a).
A similar argument holds for the left continuity of f 1 at f (b)

Exercise 19.5 (Inverse Function Theorem)

Let f : [a, b] R be continuous in [a, b] and differentiable in [a, b] with
f 0 (x) 6= 0 for all a < x < b. Let c f ([a, b]). Then there is a d [a, b] such
that f (d) = c.
(a) Let {cn } n=1 f ([a, b]) such that cn 6= c for all n 1 and limn cn = c.
Show that there is a sequence {dn } n=1 [a, b] such that

lim dn = d.
n

Hint: Exercise 14.9(b).

(b) Show that dn 6= d for all n 1.
(c) Show that
f (dn ) f (d)
lim = f 0 (d).
n dn d
Hint: Exercise 10.1.
(d) Show that f (ddnn)f
d
(d)
6= 0 for all n 1. Hint: Exercise 18.3.
(e) Show that
f 1 (cn ) f 1 (c) 1
lim = 0 .
n cn c f (d)
Thus, conclude that
1
(f 1 )0 (f (d)) =
f 0 (d)
for all d [a, b]. That is f 1 is differentiable in f ([a, b]). Hint: Exercise 10.2.

116
Solution.
(a) Since {cn }
n=1 f ([a, b]), there is a sequence {dn }n=1 [a, b] such that
f (dn ) = cn for all n 1. By the continuity of f 1 we have

lim dn = lim f 1 (cn ) = f 1 ( lim cn )) = f 1 (c) = d.

n n n

(b) If dn = d then f 1 (dn ) = f 1 (d) which implies cn = c. This contradicts

the fact that cn 6= c for all n 1.
(c) We have
f (dn ) f (d)
lim == f 0 (d)
n dn d
since f is differentiable at d. (Exercise 10.1).
(d) Since f is one-to-one (Exercise 18.3) and dn 6= d, we have f (dn ) 6= f (d).
Thus, the ratio
f (dn )f (d)
dn d
6= 0 for all n 1.
(e) We have
f 1 (cn ) f 1 (c) dn d
lim = lim
n cn c n f (dn ) f (d)
1
= lim f (dn )f (d)
n
dn d
1
=
f 0 (d)
Applying Exercise 10.2 we find
1
(f 1 )0 (c) = (f 1 )0 (f (d)) =
f 0 (d)
for all d [a, b]. That is f 1 is differentiable in f ([a, b])

Exercise 19.6
Find limx lnxx sin x+2


2x

Solution.
By LHopitals rule we have
ln x 1
lim = lim = 0.
x x x x

117
By the continuity of the sine function we have
 
x + 2
lim sin = sin = 1.
x 2x 2
Thus,   
ln x x + 2
lim sin =0
x x 2x

Exercise 19.7
Let f, g : [a, b] R be continuous on [a, b] and differentiable in a < x < b
with g 0 (x) 6= 0 for all a < x < b. Suppose that limxc f (x) = limxc g(x) =
for some a c b. Also, suppose that
f 0 (x)
lim = A.
xc g 0 (x)

Prove that
f (x)
lim = A.
xc gx(x)

Solution.
We have
1
f (x) g(x)
lim = lim 1
xc g(x) xc
f (x)
2
g 0 (x)

f (x)
= lim 0 lim
xc f (x) xc g(x)

Thus, we have
f (x) f 0 (x)
lim = lim 0 =A
xc g(x) xc g (x)

Exercise 19.8
Use LHopitals rule to evaluate limx0+ xx . Note that 00 is an undeterminate
form.

Solution.
We have
lim xx = lim+ ex ln x .
x0+ x0

118
But
ln x
lim+ x ln x = lim+ 1 = lim+ x = 0.
x0 x0 x0
x
Thus,
lim xx = elimx0+ x ln x = e0 = 1
x0+

Exercise 19.9
Let f and g be invertible differentiable functions such that

f (1) = 2; g(2) = 1; f 0 (1) = g 0 (2) = 3.

Find the derivative (f 1 g 1 )0 (1).

Solution.
We have
1 1
(f 1 g 1 )0 (1) =[f 1 ]0 (g 1 (1)) [g 1 ]0 (1) =
f 0 (f 1 (g 1 (1))) g 0 (g 1 (1))
1 1 1 1 1
= 0 1 0
= 0 0
=
f (f (2)) g (2) f (1) g (2) 9

Exercise 19.10
Let f (x) = x tan2 x for x (0, 2 ). Calculate (f 1 )0 (). Note that f ( 3 ) = .

Solution.
By the inverse function theorem we have
1 1 8
(f 1 )0 () = = = (3 + )1
f 0 (f 1 ()) f 0 ( 3 ) 3

119
Solutions to Section 20
Exercise 20.1
(a) Show that m mi (f ) Mi (f ) M.
(b) Show that m(b a) L(f, P ) U (f, P ) M (b a).

Solution.
(a) Since m is a lower bound of f in [a, b] we must have m mi (f ). Since
mi (f ) is a lower bound of f and Mi (f ) is an upper bound of f we must have
mi (f ) Mi (f ). Finally, since M is an upper bound of f in [a, b], we have
Mi (f ) M.
(b) For all 1 i n, we have m(xi xi1 ) mi (f )(xi xi1 ) Mi (f )(xi
xi1 ) M (xi xi1 ). Adding these inequalities, we obtain the desired in-
equality

Exercise 20.2
Let Q be a refinement of P. Suppose that P = {a = x0 < x1 < < xn1 <
xn = b} and Q = {a = x0 < x1 < < xi1 < z < xi < < xn = b}.
(a) Show that U (f, Q) U (f, P ).
(b) Show that L(f, P ) L(f, Q).

Solution.
Suppose first that P is a partition of [a, b] and that Q is the partition obtained
from P by adding an additional point xi1 < z < xi . The general case follows
by induction, adding one point at at time. In particular, we let

P = {a = x0 < x1 < < xn1 < xn = b}

and
Q = {a = x0 < x1 < < xi1 < z < xi < < xn = b}.
Observe that n
X
U (f, P ) = Mi (f )(xi xi1 )
i=1

and
i1
X n
X
0
U (f, Q) = Mj (f )(xj xj1 )+M (zxi1 )+M (xi z)+ Mj (f )(xj xj1 )
j=1 j=i+1

where

120
 M = sup{f (x) : x [xi1 , z]} and M 0 = sup{f (x) : x [z, xi ]}.

Since M (z xi1 ) + M 0 (xi z) Mi (xi xi1 ) we conclude that U (f, Q)

U (f, P ).
(b) Similar argument as in (a)

Exercise 20.3 Rb Rb
Suppose that f is bounded on [a, b]. Show that a f (x)dx a f (x)dx. Hint:
Exercise 20.2.

Solution.
Let P be a partition of [a, b]. For any partition Q of [a, b] we let R = P Q.
Since P R and Q R we can use Exercise 20.2 to write

L(f, P ) L(f, R) U (f, R) U (f, Q).

Since Q was arbitrary, L(f, P ) is a lower bound of SU . But the Riemann

upper integral is the largest lower bound of SU . We conclude that
Z b
L(f, P ) f (x)dx.
a

Rb
Since P was arbitrary, the above inequality says a f (x)dx is an upper bound
of SL . But the lower Riemann integral is the smallest upper bound of SL .
That is,
Z b Z b
f (x)dx f (x)dx
a a

Exercise 20.4
Consider the function f : [a, b] R defined by

2 if a x < b
f (x) =
3 if x = 3

(a) Find numbers m and M such that m f (x) M for all x [a, b]?
(b) Show that for any partition P of [a, b] we have L(f, P ) = 2(b a).
Conclude that Z b
f (x)dx = 2(b a).
a

121
 Rb Rb
(c) Show that a f (x)dx 2(b a). (d) Suppose a f (x)dx > 2(b a). Let
Rb
 = a f (x)dx 2(b a) > 0. Let Q be the partition

Q = {a = x0 < x1 < x2 < < xn = b}

Rb
such that b xn1 < . Show that U (f, Q) < a f (x)dx. Why this is impos-
sible?
(e) Is f (x) Riemann integrable? If so, what is the value of the integral
Rb
a
f (x)dx?

Solution.
(a) Since 2 f (x) 3 for all a x b, we have m = 2 and M = 3.
(b) Let P be a parition of [a, b] given by

P = {a = x0 < x1 < x2 < < xn = b}.

Note that Mi (f ) = 2 for all 1 i n 1, Mn (f ) = 3, mi (f ) = 2 for all

1 i n. Thus,
n
X
L(f, P ) = mi (f )(xi xi1 ) = 2(b a).
i=1

Since P was arbitrary, it follows that

Z b
f (x)dx = 2(b a).
a

(c) This follows from Exercise 20.3.

(d) We have U (f, Q) = n1
P
i=1 M i (f )(xi xi1 ) + Mn (f )(b xn1 ) = 2(xn1
Rb
a)+3(bxn1 ) = 2(ba)+(bxn1 ) < 2(ba)+ a f (x)dx2(ba) = inf SU ,
Rb
which contradicts the definition of infimum. We conclude that a f (x)dx =
2(b a).
Rb Rb
(e) Since a f (x)dx = a f (x)dx = 2(b a), the function f is Riemnann
integrable with Z b
f (x)dx = 2(b a)
a

122
Exercise 20.5
Consider the function f : [0, 1] R defined by f (x) = 1 if x is rational and
f (x) = 0 if x is irrational.
(a) Compute the upper Riemann integral and the lower Riemann integral.
Hint: Exercise 2.6(c).
(b) Is f Riemann integrable on [0, 1]?

Solution.
(a) Let P be a partition of [0, 1] :

P = {0 = x0 < x1 < x2 < < xn = 1}.

Clearly, Mi (f ) = 1 and mi (f ) = 0 for all 1 i n. Hence, U (f, P ) = 1 and

R1 R1
L(f, P ) = 0. It follows that 0 f (x)dx = 1 and 0 f (x)dx = 0.
(b) Since the lower Riemann sum is different from the upper Riemann sum,
the function is not Riemann integrable

Exercise 20.6
Let f : [a, b] R be a bounded function. Suppose that f is Riemann
integrable. We want to show that f satisfis the following property:

(P) > 0, there is a partition P of [a, b] such that U (f, P ) L(f, P ) < .

(a) Let  > 0 be given. Show that there is a partition P of [a, b] such that
Z b

f (x)dx < L(f, P ).
a 2

Hint: Assume the contrary and get a contradiction.

(b) Show that there is a partition Q of [a, b] such that
Z b

U (f, Q) < f (x)dx + .
a 2

(c) Let R = P Q. Use Exercise 20.2 to show that

Z b Z b
 
f (x)dx < L(f, R) U (f, R) < f (x)dx + .
a 2 a 2

(d) Show that

123
  Rb   Rb 
 
L(f, R) f (x)dx < and U (f, R) a f (x)dx <
   
 a 2 2

(e) Use the triangle inequality to show that U (f, R) L(f, R) < .

Solution.
(a) Suppose the contrary. That is,
Z b

f (x)dx L(f, P )
a 2
Rb
for all partition of [a, b]. This means that a f (x)dx 2 is an upper bound
Rb Rb
of SL . But a f (x)dx is the smallest upper bound of SL . Hence, a f (x)dx <
Rb
a
f (x)dx 2 , which is impossible.
(b) Suppose the contrary. That is,
Z b

U (f, Q) f (x)dx +
a 2
Rb
for all partition of [a, b]. This means that a f (x)dx + 2 is a lower bound of
Rb Rb
SU . But a f (x)dx is the largest lower bound of SU . Hence, a f (x)dx + 2 <
Rb
a
f (x)dx, which is impossible.
(c) We have
Z b Z b
 
f (x)dx = f (x)dx
a 2 a 2
<L(f, P ) < L(f, R) U (f, R) U (f, Q)
Z b

< f (x)dx +
a 2
Z b

= f (x)dx
a 2

(d) We conclude from (c) that

 Rb   Rb 
 
L(f, R) a f (x)dx < and U (f, R) a f (x)dx <
   
2 2

124
(e) By the triangle inequality we have U (f, R) L(f, R)  =  |U (f, R) R 
Rb Rb b
L(f, R)| = (U (f, R) a f (x)dx) + ( a f (x)dx L(f, R)) U (f, R) a f (x)dx+
   
R 
 b
 a f (x)dx L(f, R) < 2 + 2 = 


Exercise 20.7
Let f : [a, b] R be a bounded function. Suppose that f satisfies property
(P) above.
(a) Show that for each positive integer n, there is a partition Pn such that
1
U (f, Pn ) L(f, Pn ) < .
n
(b) Using (a), show that
Z b Z b
1
L(f, Pn ) f (x)dx f (x)dx < L(f, Pn ) + .
a a n
(c) Show that
Z b Z b
1
0 f (x)dx f (x)dx < .
a a n
(d) Show that
Z b Z b
f (x)dx = f (x)dx.
a a
Hint: Squeeze rule. We conclude that any bounded function that satisfies
property (P) is Riemann integrable.
Solution.
(a) This follows from property (P).
(b) We have
Z b Z b
L(f, Pn ) f (x)dx f (x)dx
a a
1
U (f, Pn ) < L(f, Pn ) +
n
(c) Using (b) we have
Z b Z b
1 1
0 f (x)dx f (x)dx < L(f, Pn ) + L(f, Pn ) = .
a a n n
(d) This follows from the squeeze rule

125
Exercise 20.8
Let f : [0, 1] R be the function f (x) = x2 . For any  > 0, choose a
partition P = {0 = x0 < x1 < < xn = 1} such that

xi xi1 < 2
for all 1 i n

Show that U (f, P ) L(f, P ) < . Hence, f is Riemann integrable.

Solution.
Pn 1
For i n we have Mi (f ) =Px2i and mi (f ) = x2i1 . Then U (f, P ) =
2 n 2
i=1 xi (xi xi1 ) and L(f, P ) = i=1 xi1 (xi xi1 ). Hence,

n
X
U (f, P ) L(f, P ) = (x2i x2i1 )(xi xi1 )
i=1
n
X
= (xi + xi1 )(xi xi1 )(xi xi1 )
i=1
Xn 
< 2 (xi xi1 )
i=1
2
n
X
(xi xi1 ) = 
i=1

Exercise 20.9
Suppose that f (x) = x for x [1, 2].
(a) Find U (f, P ) and L(f, P ). Hint: Consider a partition with equal subin-
tervals.
(b) Show that f is Riemann integrable. Hint: Exercise 20.7.
(c) Show that U (f, P ) 32 and L(f, P ) 32 .
R2
(d) Find 1 xdx.

Solution.
(a) Let P = {1 = x0 < x1 < < xn = 2} be a partition of [1, 2] with
xi = 1 + ni . Then for 1 i n we have mi (f ) = 1 + i1
n
and Mi (f ) = 1 + ni .

126
Hence,
n n  
X X i1 1
L(f, P ) = mi (f )(xi xi1 ) = 1+
i=1
n n
n
! i=1
1 1 X
= n+ (i 1)
n n i=1
  
1 1 n(n + 1)
= n+ n
n n 2
n1
=1 +
2n
and
n n  
X X i 1
U (f, P ) = Mi (f )(xi xi1 ) = 1+
i=1 i=1
n n
n
!
1 1 X
= n+ i
n n i=1
 
1 1 n(n + 1)
= n+
n n 2
n+1
=1 +
2n
(b) Let  > 0 be given. Choose n large enough so that n > 1 . Let P = {1 =
x0 < x1 < < xn = 2} be a partition of [1, 2] with xi = 1 + ni . Then
1
U (f, P ) L(f, P ) = < .
n
By Exercise 20.7, f is Riemann integrable.
(c) Since n+1
2n
is a decreasing function of n and converges to 12 , we must have
3
U (f, P ) 2 . Likewise, since n12n
is increasing and converges to 12 we must
have L(f, P ) 32 . Hence
Z b Z b
3
xdx xdx.
a
2 a

(c) From (c) we have

Z 2 Z 2 Z 2
3
xdx = xdx = xdx =
1 1 1 2

127
Exercise 20.10
Let f : [a, b] R be bounded. Let P and Q be any two partitions of [a, b].
Prove that L(f, P ) U (f, Q).

Solution.
Let R = P Q. Then L(f, R) L(f, P ) U (f, P ) U (f, R) U (f, Q)

Exercise 20.11
Let f : [a, b] R be such that m f (x) M for all x [a, b]. Prove that
Z b Z b
f (x)dx f (x)dx (M m)(b a).
a a

Solution.
Let P be a paritition of [a, b]. Then L(f, p) M (b a) and U (f, P ) m(b
Rb
a). Thus, M (b a) is an upper bound of SL so that a f (x)dx M (b a).
Rb
Likewise, m(b a) is a lower bound of SU so that m(b a) f (x)dx.
a
Rb Rb
Hence, a f (x)dx f (x)dx (M m)(b a)
a

Exercise 20.12
Let f : [a, b] R be bounded functions such that f (x) g(x) for all
x [a, b]. Prove the following:
Rb Rb
(a) a f (x)dx a g(x)dx
Rb Rb
(b) f (x)dx g(x)dx
a a

Solution.
(a) We prove (a) since (b) is similar. Let P be any partition of [a, b]. Then
Rb Rb
L(f, P ) L(g, P ) a g(x)dx. That is, a g(x)dx is an upper bound of
Rb Rb
SL (f ). This implies that a f (x)dx a g(x)dx

Exercise 20.13
Let f : [a, b] R be bounded functions. Let P be any partition of [a, b].
Prove
U (f + g, P ) U (f, P ) + U (g, P ).

128
Solution.
Let P = {a = x0 < x1 < < xn = b}. Then for all x [xi1 , xi ] we have

f (x) + g(x) Mi (f ) + Mi (g).

This implies that

Mi (f + g) Mi (f ) + Mi (g).
Hence,
U (f + g, P ) U (f, P ) + U (g, P )

Exercise 20.14
Let f : [a, b] R be Riemann integrable. Prove that there is a sequence of
Rb
partitions {Pn }
n=1 such that limn U (f, Pn ) = limn L(f, Pn ) = a f (x)dx.

Solution.
This follows from Exercise 20.7

Exercise 20.15
Consider the function f : [0, 1] R defined by f (x) = ax+b where a > 0 and
b > 0. Assume that this function is Riemann integrable. For each positive
integer n consider the partition Pn = {0 = x0 < x1 < < xn = 1} with
equal subintervals.
(a) Compute L(f,
R 1 Pn ) and Ua(f, Pn ).
(b) Show that 0 f (x)dx = 2 + b.

Solution.
(a) We have
n n
X 1 a Xi1 an1
L(f, Pn ) = (axi1 + b) = +b= +b
i=1
n n i=1 n 2 n

and n n
X 1 aXi an+1
U (f, Pn ) = (axi + b) = +b= +b
i=1
n n i=1 n 2 n
(b) We have
Z 1
a
f (x)dx = lim L(f, Pn ) = lim U (f, Pn ) = +b
0 n n 2

129
Solutions to Section 21
Exercise 21.1
Let f : [a, b] R be an increasing function on [a, b].
(a) Show that f is bounded on [a, b].
(b) Let  > 0 be given. Choose a positive integer N such that f (b)f N
(a)
< .
Let P = {a = x0 < x1 < < xn = b} be a partition of [a, b] such that
xi xi1 < N1 for all 1 i n. For each 1 i n, express Mi (f ) and mi (f )
in terms of f (x).
(c) Show that U (f, P ) L(f, P ) < . Thus, conclude that f is Riemann
integrable.

Solution.
(a) Since f is increasing, we must have f (a) x f (b) for all x [a, b]. Let
m = f (a) and M = f (b). Then f is bounded.
(b) Since f is increasing, we must have Mi (f ) = f (xi ) and mi (f ) = f (xi1 )
for 1 i n.
(c) We have
n
X
U (f, P ) L(f, P ) = [f (xi ) f (xi1 )](xi xi1 )
i=1
n
X [f (xi ) f (xi1 )]
<
i=1
N
f (b) f (a)
= <
N
It follows from Exercise 20.7 that f is Riemann integrable

Exercise 21.2
Let f : [a, b] R be a continuous function on [a, b].
(a) Show that there exist numbers m and M such that m f (x) M for
all a x b. That is, f is bounded on [a, b].
(b) Show that f is uniformly continuous on [a, b].
(c) Let  > 0. Show that there is a positive number > 0 such that if

|u v| < then |f (u) f (v)| < ba .
(d) Choose a partition P = {a = x0 < x1 < < xn = b} such that
xi xi1 < for all 1 i n. Show that for each interval [xi , xi1 ] there
exist si , ti [xi , xi1 ] such that Mi (f ) = f (ti ) and mi (f ) = f (si ). Hint:

130
Exercise 17.4.

(e) Show that Mi (f ) mi (f ) < ba for each 1 i n.
(f) Using (e), show that U (f, P ) L(f, P ) < . Hence, conclude that f is
Riemann integrable.

Solution.
(a) By Exercise 17.4 there exist x1 , x2 [a, b] such that f (x1 ) f (x) f (x2 )
for all a x b. Let m = f (x1 ) and M = f (x2 ).
(b) This follows from Exercise 14.5.
(c) This follows from the fact that f is uniformly continuous.
(d) This follows from Exercise 17.4.
 
(e) Since |ti si | < , we have |f (ti ) f (si )| < ba which implies ba <
 
f (ti ) f (si ) < ba . Thus, Mi (f )P mi (f ) = f (ti ) f (si ) < ba .
(f) We have U (f, P )L(f, P ) = ni=1 [f (ti )f (si )](xi xi1 ) < ni=1 ba 
P
(xi
xi1 ) = . We conclude that f is Riemann integrable

Exercise 21.3
Suppose f is continuous except at a point c in [a, b]. Let  > 0 be given and
consider a partition Q = {a = x0 < x1 < < xk1 < c < xk+1 < <

xn = b} such that (Q) < 12M .

(a) Prove that |xk1 xk + 1| < 6M .
(b) Show that there exist > 0 and 00 > 0 such that for all x, y [a, xk1 ]
0

with |x y| < 0 we have |f (x) f (y)| < 3(ba)


and for all x, y [xk+1 , b]
00 
with |x y| < we have |f (x) f (y)| < 3(ba) .
(c) Let P1 be a refinement of Q on [a, xk1 ] such that (P1 ) < 0 and P2 be
a refinement of P on [xk+1 , b] such that (P2 ) < 00 . Let P = P1 P2 . Then
we have
k1
X
U (f, P ) L(f, P ) = (Mi mi )(xi xi1 ) + (Mk mk )(c xk1 )
i=1
n
X
+(Mk+1 mk+1 (xk+1 c) + (Mi mi )(xi xi1 )
i=k+2

Show that
k1
X 
(Mi mi )(xi xi1 ) <
i=1
3

(Mk mk )(c xk1 ) + (Mk+1 mk+1 (xk+1 c) <
3
131
and n
X 
(Mi mi )(xi xi1 ) <
i=k+2
3

(d) Conclude that U (f, P ) L(f, P ) <  and therefore f is Riemann inte-
grable.

Solution.
  
(a) We have |xk1 xk+ | |c xk1 | + |xk+1 c| < 12M + 12M = 6M .
(b) This follows from the fact that f is uniformly continuous on [a, xk1 ] and
[xk+1 , b].
(c) We have
k1
X  
(Mi mi )(xi xi1 ) < (xk1 a) <
i=1
3(b a) 3

 
(Mk mk )(cxk1 )+(Mk+1 mk+1 (xk+1 c) < 2M (xk+1 xk ) < 2M =
6M 3
n
X  
(Mi mi )(xi xi1 ) < (b xk+1 ) <
i=k+2
3(b a) 3

(d) Taking everything together we have:

  
U (f, P ) L(f, P ) < + + = .
3 3 3
by Exercise 20.7, we conclude that f is Riemann integrable

Exercise 21.4
Suppose f is continuous except at points c1 , c2 , , cn in [a, b]. We want
to show that f is Riemann integrable on [a, b]. The proof is by induction
on n. For n = 1 the result holds by the previous exercise. Suppose that
the result holds for c1 , c2 , , cn . Suppose that f is continuous except at
c1 < c2 < < cn < cn+1 . Let  > 0. Choose > 0 small enough so that

< 8M and (cn+1 , cn+1 + ) [cn , b].
(a) Show that there is a partition P1 of [a, cn+1 ] such that U (f, P1 )
L(f, P1 ) < 4 and a partition P2 of [cn+1 , b] such that U (f, P2 ) L(f, P2 ) < 4 .
(b) Let P = P1 P2 . Show that U (f, P ) L(f, P ) < . Hence, f is Riemann
integrable on [a, b].

132
Solution.
(a) By the induction hypothesis f is Riemann integrable on [a, cn+1 ]. Since
f is continuous on [cn+1 + , b], it is integrable there. Then by Exercise 20.6
we can find a partition P1 of [a, cn+1 ] such that U (f, P1 ) L(f, P1 ) < 4
and a partition P2 of [cn+1 , b] such that U (f, P2 ) L(f, P2 ) < 4 .
(b) Let M 0 be the supremum of f and m0 the infimum of f on [cn+1 , cn+1 +
]. We have

U (f, P ) L(f, P ) =U (f, P1 ) L(f, P1 ) + (M 0 m0 )(2) + U (f, P2 ) L(f, P2 )

  
< + 2M (2) + =
4 8M 4
Thus, f is Riemann integrable on [a, b]

Exercise 21.5
Let f : [a, b] R be a increasing function on [a, b].
(a) Show that f is bounded on [a, b].
(b) Let  > 0 be given. Choose a positive integer N such that f (a)f N
(b)
< .
Let P = {a = x0 < x1 < < xn = b} be a partition of [a, b] such that
xi xi1 < N1 for all 1 i n. For each 1 i n, express Mi (f ) and mi (f )
in terms of f (x).
(c) Show that U (f, P ) L(f, P ) < . Thus, conclude that f is Riemann
integrable.

Solution.
(a) Since f is decreasing, we must have f (b) x f (a) for all x [a, b]. Let
m = f (b) and M = f (a). Then f is bounded.
(b) Since f is decreasing, we must have Mi (f ) = f (xi1 ) and mi (f ) = f (xi )
for 1 i n.
(c) We have
n
X
U (f, P ) L(f, P ) = [f (xi1 ) f (xi )](xi xi1 )
i=1
n
X [f (xi1 ) f (xi )]
<
i=1
N
f (a) f (b)
= <
N
It follows from Exercise 20.7 that f is Riemann integrable

133
Exercise 21.6
Suppose f : [a, b] R is continuous and f 0 on [a, b]. Let [c, d] [a, b].
Rb Rd
Prove that a f (x)dx c f (x)dx.

Solution.
Let P = {c = x0 < x1 < < xn = d} be a partition of [c, d]. Since
Rb Rb
f 0 We have L(f, P ) f (x)dx = a f (x)dx. Since P was arbitrary, we
Rb a Rd
can say that a f (x)dx is an upper bound of SL ([c, d]). Hence, c f (x)dx =
Rd Rb
f (x)dx a f (x)dx
c

Exercise 21.7
(a) Suppose f : [0, 1] R is continuous
R1 and f 0 on [0, 1]. Let a [0, 1] be
such that f (a) > 0. Show that 0 f (x)dx > 0.
(b)
R 1 Construct a nonnegative function f on [0, 1] such that f (0.5) > 0 but
0
f (x)dx = 0.

Solution.
(a) Let  = f (a)
2
> 0. Since f is continuous at a we can find a > 0 such that
if |x a| < then |f (x) f (a)| <  = f (a)2
. Using the triangle inequality we
f (a) R1
end up with |f (x) > 4 > 0 for all x (a , a + ). Thus, 0 f (x)dx >
R a+ R a+ f (a)
a
f (x)dx > a 4
dx = f (a)
4
> 0.
(b) Let
0 if x 6= 21

f (x) =
1 if x = 12 .
R1
Then f (0.5) = 1 > 0 and 0 f (x)dx = 0

Exercise 21.8
Suppose that f : [a, b] R is differentiable on [a, b]. Prove that f is Riemann
integrable on [a, b].

Solution.
Since f is differentiable, f is continuous; this implies f is integrable

Exercise 21.9
Let f : [a, b] R be defined by

1 if x is rational
f (x) =
1 if x is irrational

134
(a) Prove that f is not Riemann integrable on [a, b]. Hint Show that the lower
Riemann integral is different from the upper Riemann integral.
(b) Prove that |f | is Riemann integrable.

Solution.
(a) Since between any two real numbers we can find a rational number and
an irrational number, we can write U (f, P ) = b a and L(f, P ) = a b for
all P. Thus
Rb Rb
a
f (x)dx = b a and f (x)dx = a b
a

It follows that f is not integrable.

(b) Since |f (x)| = 1 for all x [a, b], |f | is continuous on [a, b] and therefore
integrable

Exercise 21.10
Suppose f is a continuous function on [a, b] and that f (x) 0 for all x [a, b].
Rb
Show that if a f (x)dx = 0, then f (x) = 0 for all x [a, b]. Hint: Assume
the contrary and get a contradiction.

Solution.
Suppose that there is an c [a, b] such that f (c) > 0. Let  = f (c) 2
> 0.
Since f is continuous at c we can find a > 0 such that if |x c| <
then |f (x) f (c)| <  = f (c)
2
. Using the triangle inequality we end up with
f (c) R1 R c+
|f (x) > 4 > 0 for all x (c , c + ). Thus, 0 f (x)dx > c f (x)dx >
R c+ f (c)
c 4
dx = f (c)
4
> 0 which is a contradiction. Hence, f (x) = 0 for all
x [a, b]

135
Solutions to Section 22
Exercise 22.1
Let f : [a, b] R be a bounded function. Suppose that lim(P )0 S(f, P ) =
A.
(a) Let  > 0. Show that there is a > 0 such that for any partition P of
[a, b] such that (P ) < we must have |S(f, P ) A| < 4 .
(b) Let Q = {a = x0 < x1 < < xn = b} be a partition of [a, b] such that
(Q) < , that is, xi xi1 < for all 1 i n. Fix 1 i n. Show

that if f (ui ) mi (f ) + 4(ba) for all xi1 ui xi then this contradicts the
definition of mi (f ).

(c) With Q as above, show that if f (vi ) Mi (f ) 4(ba) for all xi1 vi xi
then this contradicts the definition of Mi (f ).
(d) Show that for every 1 i n, there exists ui , vi [xi1 , xi ] such that
 
f (ui ) < mi (f ) + 4(ba) and f (vi ) > Mi (f ) 4(ba)
(e) Show that i=1 f (ui )(xi xi1 ) < L(f, Q)+ 4 and ni=1 f (vi )(xi xi1 ) >
Pn P
U (f, Q) 4
(f) Show that
Pn
A 4 < P 
i=1 f (ui )(xi xi1 ) < A + 4 and
n
A 4 < i=1 f (vi )(xi xi1 ) < A + 4


(g) Use (f) to show that

 
A < L(f, Q) U (f, Q) < A + .
2 2
(h) Show that U (f, Q) L(f, Q) < . That is, f is Riemann integrable.
(i) Show that Z b 
 

 f (x)dx A  < .

a
Rb
(k) Use the Squeeze rule to show that a f (x)dx = A.
Conclusion: Suppose that there is a number A such that lim(P )0 S(f, P ) =
Rb
A. Then f is Riemann integrable with a f (x)dx = A.

Solution.
(a) This follows from the hypothesis that lim(P )0 S(f, P ) = A.
(b) Let Q = {a = x0 < x1 < < xn = b} be a partition of [a, b] such that

(Q) < , that is, xi xi1 < for all 1 i n. If f (ui ) mi (f ) + 4(ba)

136
 
for all xi1 ui xi then mi (f ) + 4(ba) is a lower bound of f in [xi1 , xi ]

so by the definition of mi (f ) we must have mi (f ) + 4(ba) < mi (f ), which is
impossible.
 
(c) If f (vi ) Mi (f ) 4(ba) for all xi1 vi xi then Mi (f ) 4(ba) is
an upper bound of f on [xi1 , xi ]. By the definition of Mi (f ), we must have

Mi (f ) Mi (f ) 4(ba) , which is impossible.
(d) This follows from (b) and (c).
(e) We have
n n n
X X  X xi xi1 
f (ui )(xi xi1 ) < mi (f )(xi xi1 ) + = L(f, Q) + .
i=1 i=1
4 i=1 b a 4

Similar argument
Pn for the second part Pof the question.
n
(f) Since i=1 f (ui )(xi xi1 ) and i=1 f (vi )(xi xi1 ) are Riemann sums
with (Q) < they are P within 4 of A.
 n 
Pn we have A 2 <  i=1 f (ui)(xi xi1 ) 4 L(f, Q) U (f, Q)
(g)
i=1 f (vi )(xi xi1 ) + 4 < A + 2 .
(h) Since both U (f, Q) and L(f, Q) are inside the interval centered at A and
of length at most , we must have U (f, Q) L(f, Q) < . By Exercise ??, f
is Riemann integrable. R 
Rb  b
(i) Since L(f, Q) a f (x)dx U (f, Q), we must have  a f (x)dx A < .


(k) Let  0 and apply the Squeeze rule

Exercise 22.2

Prove that A L(f, R) < 2
and U (f, R) A < 2 .

Solution.
by Exercise ?? we have L(f, R) L(f, P ) and U (f, R) U (f, P ). Thus,
A L(f, R) A L(f, P ) < 2 and U (f, R) A U (f, P ) A < 2

Exercise 22.3
(a) For 1 i m such that L(f, Ri ) mi (zi zi1 ) 6= 0 and Mi (zi zi1 )
U (f, Ri ) prove that
L(f, Ri ) mi (zi zi1 ) < 2M and Mi (zi zi1 ) U (f, Ri ) < 2M .
(b) Use (a) and the sums above to show that
 
L(f, R) L(f, Q) < 2
and U (f, Q) U (f, R) < 2
>

137
Solution.
(a) We have

L(f, Ri ) mi (zi zi1 ) 2M (zi zi1 ) < 2M

and
Mi (zi zi1 ) U (f, Ri ) 2M (zi zi1 ) < 2M
(b) Because there are at most n 1 such terms, we obtain the bounds

L(f, R) L(f, Q) < 2nM =
2
and

U (f, Q) U (f, R) < 2nM =
2
Exercise 22.4
Use Exercise 22.2 and 22.3 to prove that

U (f, Q) < A +  and L(f, Q) > A .

Solution.
We have
 
U (f, Q) A = [U (f, Q) U (f, R)] + [U (f, R) A] < +
2 2
and
 
A L(f, Q) = [A L(f, R)] + [L(f, R) L(f, Q)] < + =
2 2
Exercise 22.5
Using the previous problem, show that

A  < S(f, Q) < A + .

That is,
|S(f, Q) A| < .

Solution.
We have A  L(f, Q) S(f, Q) U (f, Q) < A + 

138
Exercise 22.6
Suppose that f : [a, b] R is bounded and Riemann integrable. The goal of
this problem is to show that for any sequence {Pn } n=1 of partitions of [a, b]
Rb
such that limn (Pn ) = 0 we have limn S(f, Pn ) = a f (x)dx.
(a) Let  > 0. Show that there is a > 0 such that if P is a partition of [a, b]
with (P ) < we have
 Z b 
 
S(f, P ) f (x)dx  < .
 
a

(b) Show that there is a positive integer N such that if n N then (Pn ) < .
(c) Use (a) and (b) to conclude that for n N we have
 Z b 
 
S(f, Pn ) f (x)dx  < .
 
a

Hence, Z b
lim S(f, Pn ) = f (x)dx.
n a

Solution.
(a) This follows from Exercise 16.6.
(b) This follows from the definition of convergence of a sequence.
(c) If n N then (Pn ) < and this implies by (a) that
 Z b 
 
S(f, Pn ) f (x)dx < 

a

Exercise 22.7
Let f : [a, b] R be bounded and Riemann integrable. Let  > 0 be given.
Show that there is a > 0 such that for any partition P of [a, b] with (P ) <
we have
U (f, P ) L(f, P ) < .

Solution.
This follows from Exercise 22.4

Exercise 22.8
Suppose that f : [a, b] R is differentiable in [a, b] and that f 0 : [a, b] R is
Riemann integrable. Let Pn = {a = x0 < x1 < < xn = b} be a partition

139
of [a, b] such that xi xi1 = ba n
.
(a) For each 1 i n, show that there exists xi1 < ti < xi such that
f (xi ) f (xi1 ) = f 0 (ti )(xi P
xi1 ).
(b) Show that S(f , Pn ) = ni=1 f 0 (ti )(xi xi1 ) = f (b) f (a).
0

(c) Show that limn (Pn ) = 0.

Rb
(d) Show that limn S(f 0 , Pn ) = a f 0 (x)dx.
(e) Show that
Z b
f 0 (x)dx = f (b) f (a).
a

Solution.
(a) This P follows from the Mean Value Theorem.
(b) We ni=1 f 0 (ti )(xi xi1 ) = (f (x1 ) f (a)) + (f (x2 ) f (x1 )) + +
(f (xn1 ) f (xn2 ) + (f (b) f (xn1 )) = f (b) f (a).
(c) We have limn (Pn ) = limn ba n
= 0.
(d) This follows from the previous exercise.
(e) We have
Z b
f 0 (x)dx = lim S(f 0 , Pn ) = lim (f (b) f (a)) = f (b) f (a)
a n n

Exercise 22.9 (Fundamental Theorem of Calculus)

Suppose that f : [a, b] R is continuous and let F : [a, b] R be a
differentiable function such that F 0 (x) = f (x) for all a x b. Show that
Z b
f (x)dx = F (b) F (a).
a

The function F (x) is called an antiderivative of f.

Solution.
Using the previous problem we see that
Z b Z b
f (x)dx = F 0 (x)dx = F (b) F (a)
a a

140
Solutions to Section 23
Exercise 23.1
Let f, g : [a, b] R be Riemann integrable functions and , be real num-
bers. Let  > 0.  Rb 
(a) Show that there is a 1 > 0 such that if (P ) < 1 then S(f, P ) a f (x)dx <
 

||+||
.
 Rb 
(b) Show that there is a 2 > 0 such that if (P ) < 2 then S(g, P ) a g(x)dx <
 

||+||
.
(c) Show that there is a > 0 such that if (P ) < then
  Z b Z b 
 
S(f + g, P ) f (x)dx + g(x)dx  < .
 
a a

We conclude that f + g is Riemann integrable and

Z b Z b Z b
(f (x) + g(x))dx = f (x)dx + g(x)dx.
a a a

Solution.
(a) Follows from Exercises 22.2 - 22.5.
(b) Follows from Exercises 22.2 - 22.5.
(c) Let = min{1 , 2 }. Suppose that (P ) < . Then (P ) < 1 and
(P ) < 2 . Moreover, we have
  Z b Z b 
 
S(f + g, P ) f (x)dx + g(x)dx 
 
a a
  Z b Z b 
 
= S(f, P ) + S(g, P ) f (x)dx + g(x)dx 
a a
 Z b   Z b 
   
|| S(f, P ) f (x)dx + || S(g, P ) g(x)dx
a a
 
<|| + || = .
|| + || || + ||

By Exercise 22.1, f + g is Riemann integrable and

Z b Z b Z b
(f (x) + g(x))dx = f (x)dx + g(x)dx
a a a

141
Exercise 23.2
Let f, g : [a, b] R be Riemann integrable functions such that f (x) g(x)
for all x [a, b].
(a) Show that for any partition P of [a, b] we have L(f, P ) L(g, P ).
Rb Rb
(b) Show that f (x)dx g(x)dx.
R ba R ba
(c) Show that a f (x)dx a g(x)dx.

Solution.
(a) Since f (x) P g(x) for all x [a, b], wePhave mi (f ) mi (g) for all
1 i n. Thus, ni=1 mi (f )(xi xi1 ) ni=1 mi (g)(xi xi1 ), that is,
L(f, P ) L(f, g).
(b) From (b), we have
Z b
L(f, P ) L(g, P ) g(x)dx
a
Rb
for all partitions P of [a, b]. Thus, g(x)dx is an upper bound for SL =
a
Rb
{L(f, P ) : P a partition of [a, b}. But a f (x)dx is the smallest upper bound
so that Z b Z b
f (x)dx g(x)dx.
a a

(c) This follows from the fact that

Rb Rb Rb Rb
a
f (x)dx = a
f (x)dx and a
g(x)dx = a
g(x)dx

Exercise 23.3
Let f : [a, b] R be a Riemann integrable function such that m f (x) M
for all x [a, b].
(a) Show that m(b a) L(f, P ) U (f, P ) M (b a) for any partition
P of [a, b].
Rb Rb
(b) Show that a f (x)dx = a f (x)dx M (b a).
Rb Rb
(c) Show that m(b a) a f (x)dx = a f (x)dx.
Rb
Conclusion: m(b a) a f (x)dx M (b a).

Solution.
(a) Let P = {a = x0 < x1 < < xn = b}. Then
m(xi xi1 ) mi (f )(xi xi1 ) Mi (f )(xi xi1 ) M (xi xi1 ).

142
Summing from i = 1 to i = n to obtain

m(b a) L(f, P ) U (f, P ) M (b a)

Rb
(b) M (b a) is an upper bound of SL . But a f (x)dx is the smallest upper
bound so that Z b
f (x)dx M (b a).
a

Since f is Riemann integrable, we have

Z b Z b
f (x)dx = f (x)dx M (b a).
a a

Rb
(c) m(b a) is a lower bound of SU . But a
f (x)dx is the largest lower bound
so that Z b
m(b a) f (x)dx.
a
Since f is Riemann integrable, we have
Z b Z b
m(b a) f (x)dx = f (x)dx
a a

Exercise 23.4
Let f : [a, b] R be a Riemann integrable function and a < c < b.
(a) Let  > 0. Show that there is a partition P of [a, b] such that U (f, P )
L(f, P ) < .
(b) Let Q = P {c}, Q1 = Q [a, c], and Q2 = Q [c, b]. That is, Q is
partition of [a, b], Q1 is a partition of [a, c], and Q2 is a partition of [c, b].
Show that

[U (f, Q1 ) L(f, Q1 )] + [U (f, Q2 ) L(f, Q2 )] < .

Rc
(c) Show that U (f, Q1 ) L(f, Q1 ) < . Thus, by Exercise 21.1, a f (x)dx
exists and is finite. Rb
(d) Show that U (f, Q2 ) L(f, Q2 ) < . Thus, by Exercise 21.1, c f (x)dx
exists and is finite.

143
Solution.
(a) This follows from Exercise 21.1.
(b) We have

[U (f, Q1 ) L(f, Q1 )] + [U (f, Q2 ) L(f, Q2 )] =[U (f, Q1 ) + U (f, Q2 ]

[L(f, Q1 )] + L(f, Q2 )]
=U (f, Q) L(f, Q)
U (f, P ) L(f, P ) < 

(c) Since U (f, Q1 ) L(f, Q1 ) > 0 we must have U (f, Q1 ) L(f, Q1 ) < . By
Exercise 21.1, f is Riemann integrable in [a, c].
(d) Since U (f, Q2 ) L(f, Q2 ) > 0 we must have U (f, Q2 ) L(f, Q2 ) < . By
Exercise 21.1, f is Riemann integrable in [c, b]

Exercise 23.5
Let f : [a, b] R be a Riemann integrable function and a < c < b. Let  > 0.
(a) Show that there is a 1 > 0 Rsuch that if P1 is a partition of [a, c] such
c
that (P1 ) < 1 then S(f, P1 ) a f (x)dx < 2 .
(b) Show that there is  a 2 > 0 Rsuch that if P2 is a partition of [c, b] such
b
that (P2 ) < 2 then S(f, P2 ) c f (x)dx < 2 .
 

(c) Let P = P1 P2 . Then P is a partition of [a, b]. Show that there is > 0
such that (P ) < and
 Z c Z b 
 
S(f, P ) f (x)dx + f (x)dx  < .
 
a c

That is, Z b Z c Z b
f (x)dx = f (x)dx + f (x)dx.
a a c

Solution.
(a) By the previous exercise, f is Riemann integrable in [a, c]. Now, the result
follows from Exercises 22.2 - 22.5.
(b) Similar to (a).

144
(c) Let = min{1 , 2 }. Then (P ) < . Also, we have
 Z c Z b 
 
S(f, P ) f (x)dx + f (x)dx 
 
a c
 Z c   Z b 
   
S(f, P1 )
 f (x)dx + S(f, P2 )
  f (x)dx
a c
 
< + =
2 2
Now the result follows from Exercise 22.1

Exercise 23.6
Let f : [a, b] R be continuous. Use the Intermediate Value Theorem to
prove the existence of a number c [a, b] such that
Z b
f (x)dx = (b a)f (c).
a

The number f (c) is called the average value of f on [a, b].

Solution. Rb
We know that m(b a) a f (x)dx M (b a) (Exercise 23.3). Let
1
Rb
d = ba a
f (x)dx. Then d [m, M ]. By the IVT there is a number c [a, b]
such that f (c) = d. This implies that
Z b
f (x)dx = (b a)f (c)
a

Exercise 23.7 Rb
Suppose that f and g are continuous function on [a, b] such that a f (x)dx =
Rb
a
g(x)dx. Prove there is a c [a, b] such that f (c) = g(c).

Solution.
Let h(x) = f (x) g(x). Then h is continuous on [a, b]. By the previous
1
Rb Rb
exercise, there is a c [a, b] such that ba a
h(x)dx = h(c). But a
h(x)dx =
0. Thus, h(c) = 0 or f (c) = g(c)

Exercise 23.8
(a) For any set S, one can see that M (f, S) m(f, S) = sups,tS |f (s) f (t)|.

145
Let f be a function defined on a set S. Show that M (|f |, S) m(|f |, S)
M (f, S) m(f, S).
(b) Suppose that f : [a, b] R is Riemann integrable. Show that |f is also
Riemann integrable.

Solution.
(a) For any s, t S we have

||f (s)| |f (t)|| |f (s) f (t)|.

It follows that

M (|f |, S)m(|f |, S) = sup ||f (s)||f (t)|| sup |f (s)f (t)| = M (f, S)m(f, S).
s,tS s,tS

(b) Let  > 0 be given. Then there is a partition P = {a = x0 < x1 < <
xn = b} such that U (f, P ) L(f, P ) < . From part (a) we have that for
1in
Mi (|f |) mi (|f |) Mi (f ) mi (f ).
Hence,
U (|f |, P ) L(|f |, P ) < U (f, P ) L(f, P ) < .
That is, |f | is integrable

Exercise 23.9
Let f : [a, b] R be defined by

1 if x Q
f (x) =
1 if x 6 Q
Rb Rb
(a) Compute f (x)dx and a f (x)dx.
a
(b) Is f Riemann integrable?
(c) Show that |f | is Riemann integrable.

Solution.
Let P = {a = x0 < x1 < < xn = b} be any partition of [a, b]. Then
mi = 1 and Mi = 1 for all 1 i n. Hence, L(f, P ) = a b and
U (f, P ) = b a. It follows that
Rb Rb
f (x)dx = a b and a
f (x)dx = b a.
a

146
(b) It follows from (a) that f is not Riemann integrable.
(c) |f |(x) = 1 for all x [a, b]. Since |f | is a continuous function, |f | is
integrable on [a,b]

Exercise 23.10
Let f : [a, b] R be Riemann integrable with |f (x)| M for all x [a, b].
(a) Prove that |f 2 (x) f 2 (y)| 2M |f (x) f (y)| for all x, y [a, b] where
f 2 (x) = (f (x))2 .
(b) Let  > 0. Show that there is a partition P of [a, b] such that

U (f, P ) L(f, P ) < .
2M
(c) Prove that U (f 2 , P ) L(f 2 , P ) < . That is, f 2 is Riemann integrable.

Solution.
(a) We have |f 2 (x)f 2 (y)| = |f (x)+f (y)||f (x)f (y)| (|f (x)|+|f (y)|)|f (x)
f (y)| 2M |f (x) f (y)|.
(b) This follows from Exercise 20.6.
(c) Let P = {a = x0 < x1 < < xn = b}. We have
n
X
2 2
U (f , P ) L(f , P ) = sup |f 2 (s) f 2 (t)|(xi xi1 )
i=1 s,t[xi1 ,xi ]
X n
2M sup |f (s) f (t)|(xi xi1 )
i=1 s,t[xi1 ,xi ]

=2M (U (f, P ) L(f, P )) < 2M =
2M
Exercise 23.11
Let f, g : [a, b] R be two Riemann integrable functions.
(a) Show that
1
f g = [(f + g)2 f 2 g 2 ].
2
(b) Prove that f g is Riemann integrable.

Solution.
(a) Trivial algebra.
(b) Since f and g are integrable so are the functions f 2 , g 2 , and (f + g)2
according to the previous problem. Now the result follows from Exercise
23.1

147
Solutions to Section 24
Exercise 24.1
Suppose that f : [a, b] [c, d] is a Riemann integrable function on [a, b] and
that g : [c, d] R is continuous (and hence integrable by Exercise ??).
(a) Show that the set {|g(x)| : x [c, d]} is bounded. Hence, by the Com-
pleteness Axiom of R there exists K > 0 such that K = sup{|g(x)| : x
[c, d]}.
(b) Let  > 0. Chosse 0 so that 0 < ba+2K
. Show that there is a < 0 such
that if |s t| < , where s, t [c, d], then |g(s) g(t)| < 0 .
(c) Show that there is a partition P = {a = x0 < x1 < < xn = b} of [a, b]
such that U (f, P ) L(f, P ) < 2 .
(d) Let A = {1 i n : Mi (f ) mi (f ) < }. Show that if i A then
Mi (g f ) mi (g f )| < 0 . P
(e) Let B = {1P i n : Mi (f )mi (f ) }. Show that iB (xi xi1 ) <
2 and hence iB (xi xi1 ) < .
(f) Show that for all 1 i n we have Mi (g f ) mi (g f ) < 2K. Hint:
Use Exercise 15.8 and the triangle inequality.
(g) Use (d) (e) and (f) to show that U (g f, P ) L(g f, P ) < . Hence, by
Exercise 20.7, g f : [a, b] R is Riemann integrable.

Solution.
(a) This follows from Exercise 15.3.
(b) This is just the uniform continuity of g. See Exercise 14.5(d).
(c) This follows from Exercise 20.6 since f is Riemann integrable on [a, b].
(d) Let i A and s, t [xi1 , xi ]. Since Mi (f ) mi (f ) < , we must have
|f (s) f (t)| < . Now by (b) we have |g(f (s)) g(f (t))| < 0 and this implies
that Mi (g f )P mi (g f )| < 0 . P
(e) We have iB (xi xi1 ) < iB [Mi (f )mi (f )](xi xi1 ) U (f, P )
L(f, P ) < 2 . Divide both sides by to obtain the required result.
(f) Note that by Exercise 15.8 we have Mi (g f ) g([c, d]) and mi (g f )
g([c, d]). Hence, Mi (g f ) mi (g f ) |Mi (g f ) mi (g f )| |Mi (g
f )| + |mi (g f )| < K + K = 2K.

148
(g) We have
n
X
U (g f, P ) L(g f, P ) = [Mi (g f ) mi (g f )](xi xi1 )
i=1
X
= [Mi (g f ) mi (g f )](xi xi1 )
iA
X
+ [Mi (g f ) mi (g f )](xi xi1 )
iB
X X
<0 (xi xi1 ) + 2K (xi xi1 )
iA iB
Xn
<0 (xi xi1 ) + 2K
i=1
=0 (b a) + 2K
<0 (b a) + 2K0
=0 (b a + 2K) < 

Thus, by Exercise 20.7, g f is Riemann integrable

Exercise 24.2
Let f, g : [a, b] R be Riemann integrable and bounded such that |f (x)|
M1 and |g(x)| M2 for all x [a, b].
(a) Find a positive constant M such that |f (x)| M and |g(x)| M. Thus,
f ([a, b]) [M, M ] and g([a, b]) [M, M ]
(b) Consider the continuous function h : [2M, 2M ] R given by h(x) = x2 .
Show that (f + g)2 and (f g)2 are Riemann integrable on [a, b]. Hint: Note
that h (f + g) = (f + g)2 and h (f g) = (f g)2 .
(c) Show that f g is Riemann integrable on [a, b].

Solution.
(a) Let M = M1 + M2 .
(b) This follows from Exercise 24.4.
(c) Since f g = 14 [(f + g)2 (f g)2 ], by Exercise 23.1 we conclude that
f g is Riemann integrable on [a, b]

Exercise 24.3
Let f : [a, b] R be Riemann integrable and bounded such that |f (x)| M

149
for all x [a, b].
(a) consider the continuous function g : [M, M ] R defined by g(x) = |x|.
Show that |f | is Riemann integrable on [a, b].
(b) Using the fact that |f (x)| f (x) |f (x)| for all x [a, b], show that
Z b Z b Z b
|f (x)|dx f (x)dx |f (x)|dx.
a a a

Hence, show that Z b  Z b

 
 f (x)dx  |f (x)|dx.
 
a a

Solution.
(a) This follows from Exercise 24.4.
(b) This follows from Exercise 23.2. The last equality follows from Exercise
1.14

Exercise 24.4 (Integration by Parts)

Let f, g : [a, b] R be continuous and f 0 , g 0 : [a, b] R be Riemann
integrable.
(a) Show that f and g are Riemann integrable on [a, b].
(b) Show that f 0 g and f g 0 are Riemann integrable on [a, b].
Rb Rb
(c) Show that a f 0 gdx + a f g 0 dx = (f g)(b) (f g)(a). Hint: Use product
rule and Exercise 22.8.

Solution.
(a) This follows from Exercise 21.2.
(b) This follows from Exercise 24.5.
(c) By the product rule we have (f g)0 = f 0 g + f g 0 . Hence, using Exercise 22.8
we can write
Z b Z b Z b
0 0
f gdx + f g dx = (f g)0 dx = (f g)(b) (f g)(a)
a a a

or Z b Z b
0
f gdx = (f g)(b) (f g)(a) f g 0 dx
a a

150
Exercise 24.5
Consider the function f : [0, 1] R defined by

1 if 0 < x 1
f (x) =
0 if x = 0.
R1
Show that f is Riemann integrable on [0, 1]. What is the value of 0
f (x)dx?

Solution.
Let P be a parition of [0, 1] given by
P = {0 = x0 < x1 < x2 < < xn = 1}.
Note that Mi (f ) = 1 for all 1 i n, m1 (f ) = and mi (f ) = 1 for all
1 i n 1. Thus,
n
X
U (f, P ) = Mi (f )(xi xi1 ) = 1.
i=1

Since P was arbitrary, it follows that

Z 1
f (x)dx = 1.
0
R1 R1
We know from Exercise 20.3 that 0 f (x)dx 1. Suppose that 0 f (x)dx < 1.
R1
Let  = 1 0 f (x)dx > 0. Let Q = {0 = x0 < x1 < < xn = 1} be
a partition of [0, 1] such that x1 R> . Then L(f, Q) = m1 (f )(x1 x0 ) +
Pn1 1
i=2 mi (f )(xi xi1 ) = 1 x1 < 0 f (x)dx which contradicts the definition
R1
of supremum. We conclude that 0 f (x)dx = 1.
R1 R1
Since 0 f (x)dx = 0 f (x)dx = 1, the function f is Riemnann integrable with
Z 1
f (x)dx = 1
0

Exercise 24.6
Consider the function g : [0, 1] R defined by

1 if x = 0 or x = 1
1 p
g(x) = if x = q is rational with p and q > 0 in lowest terms
q
0 if x is irrational.

151
(a) Let  > 0 and 0 = min{0.5, }. Thus, 0 < 0 0.5 and 0 < 0 . Show
0
that there is a finite number of rationals in [0, 1] such that g(x) 2 . Denote
the rationals by {r0 , r1 , , rn } where r0 = 0 and rn = 1.
(b) Define the partition Q = {0 = x0 < x1 < x2 < < x2n < x2n+1 = 1}
0
where x0 = 0; x1 < r1 with x1 < 2(n+1) ; x1 < x2 < r1 < x3 with x3 x2 <
0 0
2(n+1)
; ; x2n2 < rn1 < x2n1 with x2n1 x2n2 < 2(n+1)
; x2n1 < x2n <
0 0
1 with 1 x2n < and x2n+1 = 1. Show that U (g, Q) <  . Hint: Note
2(n+1)
that the sum involves intervals containing ri0 s and those that do not.
(c) Show that L(g, Q) = 0. Hint: Exercise 2.6.
(d) Using (b) and (c) show that U (g, Q) L(g, Q) < . Thus, g is Riemann
integrable. R1
(e) What is the value of the integral 0 g(x)dx?

Solution.
0 0
(a) Let x = pq (0, 1) such that g(x) 2 . Then 1q 2 which implies that
0 < q 20 . This shows that there are only a finite number of positive integers
0
q that satisfy this inequality. Moreover, g(0) = g(1) > 2 . Lets denote these
rationals by {r0 , r1 , , rn } where r0 = 0 and rn = 1.
(b) The terms of U (g, Q) consists of two types of intervals: The intervals
[xi1 , xi ] not containing ri s and in this case we have Mi (g)(xi xi1 )
0
2
(xi xi1 ) and there are n such intervals. The second type consists of
0
those intervals containing ri s and in this case Mi (g)(xi xi1 ) < 2(n+1) and
there are n + 1 such intervals. Thus,
n
X n
X
U (g, Q) = Mi (g)(x2i+1 x2i ) + Mi (g)(x2i x2i1 )
i=0 i=1
n n
X 0 X
< (x2i+1 x2i ) + (x2i x2i1 )
i=0
2 i=1
0 0
<(n + 1) + = 0 .
2(n + 1) 2

(c) By Exercise 2.6, we have mi (g) = 0 for all i. Hence, L(g, Q) = 0. There-
fore, U (g, Q) L(g, Q) < .
(d) By Exercise 20.7 and (c) , we conslude that g is Riemann R 1 integrable.
(e) Since L(g, P ) = 0 for all partitions P of [0, 1] we find 0 g(x)dx = 0 =
R1
0
g(x)dx

152
Exercise 24.7
Consider the functions f and g introduced in the previous two exercises. Let
h(x) = (f g)(x).
(a) Write explicitly the formula of h(x) as a piecewise defined function.
(b) Show that h is not Riemann integrable on [0, 1].

Solution.
(a) We have
 
1 if 0 < g(x) 1 1 if x is rational
h(x) = =
0 if g(x) = 0 0 if x is irrational.

(b) This follows from Exercise 20.5

Exercise 24.8
Let f, g : [a, b] R be Riemann integrable.
(a) Show that max{f (x), g(x)} = |f (x)g(x)|+f
2
(x)+g(x)
.
(b) Show that the function max{f (x), g(x)} is Riemann integrable.

Solution.
(a) If f (x) g(x) then max{f (x), g(x)} = f (x) and |f (x)g(x)|+f
2
(x)+g(x)
=
f (x)g(x)+f (x)+g(x)
2
= f (x). Similar argument when f (x) < g(x).
(b) All functions |f g|, f, and g are Riemann integrable so the combination
max{f (x), g(x)} is also Riemann integrable

Exercise 24.9
Let f, g : [a, b] R be Riemann integrable.
(a) Show that min{f (x), g(x)} = f (x)+g(x)|f
2
(x)g(x)|
.
(b) Show that the function min{f (x), g(x)} is Riemann integrable.

Solution.
Similar to the previous exercise

153
Solutions to Section 25
Exercise 25.1
Let f and F as defined in Definition 24. Let M be such that |f (x)| M for
all x [a, b]. Fix c in [a, b].
(a) Show that for any x [a, b] we have
Z x
M (x c) f (t)dt M (x c).
c
Hence, we can write
x
Z 
 

 f (t)dt M |x c|.
c
Hint: Exercise 23.2.
(b) Let  > 0 and = M . Show that for any x [a, b] such that |x c| <
we must have |F (x) F (c)| < . Hence, F is continuous at c. Since c was
arbitrary in [a, b], we conclude that F is continuous on [a, b].
Solution.
(a) Since M f (t) M for all t [a, b] we can use Exercise 23.2 to obtain
Z x Z x Z x
M dt f (t)dt M dt
c c c
which implies Z x
M (x c) f (t)dt M (x c)
c
or Z x

 

 f (t)dt M |x c|.
c
(b) We have
Z x Z c 
 
|F (x) F (c)| = 
 f (t)dt f (t)dt
Za a Zax 
 
= f (t)dt + f (t)dt
Zc x  a
 
=  f (t)dt M |x c|
c
<M = 
Hence, F is continuous at c. Since c was arbitrary in [a, b], we conclude that
F is continuous on [a, b]

154
Exercise 25.2
Let f and F as above. Suppose furthermore that f is continuous at c [a, b].
(a) Show that

F (c + h) F (c) 1 c+h
Z
f (c) = [f (t) f (c)]dt.
h h c
(b) Show that F 0 (c) exists and is equal to f (c).

Solution.
(a) We have

F (c + h) F (c) 1 c+h 1 c+h

Z Z
f (c) = f (t)dt dt
h h c h c
1 c+h
Z
= [f (t) f (c)]dt.
h c
(b) Let  > 0. By the continuity of f at c we can find a > 0 such that if
|x c| < then |f (x) f (c)| < . Choose h such that |h| < . Then for any t
between c and c + h we have |t c| < and therefore |f (t) f (c)| < . Thus,
  Z c+h 
 F (c + h) F (c)  1  
 f (c)=
 |h| 
 [f (t) f (c)]dt
 h c

1
< |h| = .
|h|

It follows that F 0 (c) exists and is equal to f (c)

Exercise 25.3
Suppose that f : [a, b] R is differentiable on [a, b] and f 0 continuous on
[a, b].
(a) Show that f 0 is Riemann integrableR x on [a, b].
(b) Define F : [a, b] R by F (x) = a f (t)dt. Show that F 0 (x) = f 0 (x) for
0

all x [a, b].

(c) Show that F (x) = f (x) f (a) for all x [a, b].
(d) Use (c) to show that
Z b
f 0 (x)dx = f (b) f (a).
a

155
Solution.
(a) This follows from Exercise 21.2.
(b) This follows from Exercise 25.2.
(c) From Exercise 18.9 we have F (x) = f (x) + C for all x [a, b], where C
is a constant. Letting x = a we find 0 = F (a) = f (a) + C or C = f (a).
Thus, F (x) = f (x) f (a).
(d) The result follows by letting x = b in (c)
Exercise 25.4
Suppose that f : [a, b] R is continuous on [a, b] and g : [c, d] [a, b] is
differentiable on [a, b]. Define F : [c, d] R by
Z g(x)
F (x) = f (t)dt.
a
(a) Show that f is Riemann integrable R x on [a, b].
(b) Define G : [a, b] R by G(x) = a f (t)dt. Show that G is differentiable
and G0 (x) = f (x) for all x [a, b].
(c) Write F in terms of G and g. Show that F is differentiable on [c, d] with
F 0 (x) = f (g(x)) g 0 (x).
Solution.
(a) This follows from Exercise 21.2.
(b) This follows from Exercise 25.2.
(c) We have F (x) = (G g)(x). Since both G and g are differentiable, by
Exercise 16.10, we find that F is differentiable with derivative
F 0 (x) = G0 (g(x)) g 0 (x) = f (g(x)) g 0 (x)
Exercise 25.5 (Mean Value Theorem for Integrals)
Let f : [a, b] R be continuous.
(a) Show that f is Riemann integrable on [a, b].
(b) Define F : [a, b] R by
Z x
F (x) = f (t)dt.
a
0
Show that F is differentiable with F (x) = f (x).
(c) Show that there is a < c < b such that F (b) F (a) = F 0 (c)(b a).
(d) Use (c) to show that
Z b
f (x)dx = f (c)(b a).
a

156
Solution.
(a) This follows from Exercise 21.2.
(b) This follows from Exercise 25.2.
(c) This follows from the Mean Value Theorem (Exercise 18.1).
Rb
(d) Since F (a) = 0, we have F (b) = F (b) F (a) = a f (x)dx. Also, F 0 (c) =
f (c). Hence,
Z b
f (x)dx = f (c)(b a)
a

Exercise 25.6 (Change of Variables Formula)

Let : [a, b] [c, d] be differentiable with continuous derivative and such
that (a) = c, (b) = d. Let f : [c, d] R be continuous.
(a) Show that the functions
Rx f and (f ) 0 are Riemann integrable.
(b) Define F (x) = c f (t)dt. Show that F is differentiable with F 0 (x) = f (x)
for all x [c, d]. Rx
(c) Define G(x) = a f ((t))0 (t)dt. Show that G is differentiable with
G0 (x) = f ((x))0 (x) for all x [a, b].
(d) Show that F is differentiable on [a, b] with (F )0 (x) = G0 (x) for all
x [a, b]. Hint:Exercise ??.
(e) Use (d) and Exercise 18.9 to show that (F )(x) = G(x) for all x [a, b].
(f) Use (e) to show that
Z b Z d
0
f ((x)) (x)dx = f (x)dx.
a c

Solution.
(a) Since f, , and 0 are continuous, they are Riemann integrable by Exercise
21.2. By Exercise 24.1, f is Riemann integrable. Also, f is continuous
being the composition of two continuous functions. Since f : [a, b] R,
by Exercise 15.3 , f is bounded. Since 0 : [a, b] R, 0 is also bounded.
Thus, by Exercise 24.2, the product (f ) 0 is integrable.
(b) Since f is continuous, from Exercise 25.2 we conclude that F is differen-
tiable with F 0 (x) = f (x) for all x [c, d].
(c) Since (f ) 0 is continuous, by Exercise 25.2, G is differentiable with
G0 (x) = f ((x))0 (x) for all x [a, b].
(d) Since both F and are differentiable, the composition F is also
differentiable (Exercise 18.1) with derivative

(F )0 (x) = F 0 ((x)) 0 (x) = f ((x)) 0 (x) = G0 (x).

157
(e) Using (d) and Exercise 10.1 we can write (F )(x) = G(x) + C for
all x [a, b], where C is a constant. In particular, letting x = c we find
0 = F (c) = (F )(a) = G(a) + C = C. Thus, (F )(x) = G(x) for all
x [a, b].
(f) This follows by letting x = b in (e)

Exercise 25.7
Find the derivative of
Z x
F (x) = cos (t2 )dt.
1

Solution.
From Exercise 25.4 with g(x) = x and f (x) = cos x2 we find
cos x
F 0 (x) = f (g(x)) g 0 (x) =
2 x

Exercise 25.8 (Mean Value Theorem for Monotone Functions)

Let f : [a, b] R be increasing on [a, b].
(a) Show that f is Riemann integrable on [a, b].
(b) Define g : [a, b] R by g(x) = f (a)(x a) + f (b)(b x). Show that g is
continuous on [a, b].
Rb
(c) Show that g(b) a f (x)dx g(a).
(d) Show that there is c [a, b] such that
Z b
f (x)dx = f (a)(c a) + f (b)(c b).
a

Solution.
(a) This follows from Exercise 21.1.
(b) This follows from the fact that g is a combination of continuous functions.
(c) Since a x b and f is increasing, we find f (a) f (x) f (b). Integrate
each side from a to b we find
Z b
f (a)(b a) f (x)dx f (b)(b a)
a

or Z b
g(b) f (x)dx g(a).
a

158
(d) By the Intermediate value Theorem we can find a c [a, b] such that
Z b
g(c) = f (x)dx
a

or Z b
f (x)dx = f (a)(c a) + f (b)(c b)
a

Exercise 25.9 R3
Use change of variables to evaluate 1 (3x + 1)100 dx.

Solution.
Let f (x) = x100 and (x) = 3x + 1. Then
Z 3 Z 10
10
100 100 3 101 
(3x + 1) dx = 3 d = 
1 4 101 3

Exercise 25.10
Find the smallest positive critical point of
Z x
3
F (x) = cos (t 2 )dt.
0

Solution.
23
3
We have F 0 (x) = cos (x 2 ), so the smallest positive critical point is x =
2

Exercise 25.11
Suppose f : R R is continuous at a R. Find
Z x
1
lim f (t)dt.
xa x a a

Solution. R
x
Let F (x) = a f (t)dt. Then F is differentiable at a with F 0 (a) = f (a). Thus,
Z x
1 F (x) F (a)
lim f (t)dt = lim = F 0 (a) = f (a)
xa x a a xa xa

159
Exercise 25.12
Let f : R R be continuous and A, B : R R be differentiable functions.
Define g : R R by
Z B(x)
g(x) = f (t)dt.
A(x)

Prove that g is differentiable and find a formula for g 0 (x).

Solution. Rx
Define F (x) = 0 f (t)dt. Since f is continuous, the First Fundamental Theo-
rem of Calculus shows that F is differentiable everywhere and F 0 (x) = f (x).
Note that
Z B(x) Z A(x)
g(x) = f (t)dt f (t)dt = F (B(x)) F (A(x)).
0 0

Since A, B, F are differentiable, so is g. By the chain rule,

g 0 (x) = F 0 (B(x))B 0 (x) F 0 (A(x))A0 (x) = f (B(x))B 0 (x)0 f (A(x))A0 (x)

Exercise 25.13
Suppose f : R R is continuous at 2 and f (2) = 4. Find
Z x
1
lim xf (t)dt.
x2 x 2 2

Solution.
We have
Z x Z x
1 1
lim xf (t)dt = lim x f (t)dt = 2f (2) = 8
x2 x 2 2 x2 x2 2

Exercise 25.14
cos 2x3
Use a definite integral to define a function F (x) having derivative
1+x4
for

all x and satisfying F ( 3 2) = 0.

Solution.
The answer is x
cos 2t3
Z
F (x) =
dt
3
2 1 + t4

160
Solutions to Section 26
Exercise 26.1
Show that the series 1
P
n=1 n(n+1) converges to 1. Hint: Show that for each
1
n 1 we have Sn = 1 n+1 .
Solution.
Using partial fractions we can write
1 1 1
= .
n(n + 1) n n+1
Thus,
S1 = 1 12
S2 = (1 12 ) + ( 21 13 ) = 1 13
S3 = S2 + ( 13 14 ) = (1 31 ) + ( 13 41 ) = 1 1
4
..
.
1
Sn = 1 n+1
.
It follows that limn Sn = 1.
Exercise 26.2
Is the series n
n=1 (1) convergent or divergent?

Solution.
The series n
n=1 (1) diverges since the sequence of partial sums alternates
between the values 1 and 0.
Exercise 26.3P

Suppose that i=1 an = L. Show that limn an = 0. Hint: Note that
Sn+1 Sn = an .
Solution.
We know that Sn = a1 + a2 + + an and Sn+1 = a1 + a2 + + an + an+1 =
Sn + an so it follows that Sn+1 Sn = an . Suppose that the series converges
to a number L. Then limn Sn = limn Sn+1 = L. Thus, limn an =
limn (Sn+1 Sn ) = L L = 0.
Exercise 26.4
Consider the series ni=1 log n+1
P

n
.
(a) Show that limn an = 0.
(b) Show that limn Sn = . Hence, the series is divergent.

161
Solution.
(a) We have limn an = limn log 1 + n1 = log 1 = 0.

(b) We have that Sn = log (n + 1). Hence, limn Sn = limn log (n + 1) =

so that the given series is divergent

Exercise 26.5
Consider the sequence {rn } n=1 .
(a) Show that if r = 1 the sequence is divergent.
(b) Show that if |r| > 1, i.e. r < 1 or r > 1, the sequence is divergent.
(c) Show that if |r| < 1, the sequence is convergent.

Solution.
(a) This follows from Exercise .
(b) |r| > 1 implies r > 1 or r < 1. Suppose first that r > 1. Let  > 0. Let

N be a positive integer greater than r1 . Then for n N we have

rn = (1 + (r 1))n
1 + n(r 1)(by the binomial f ormula)
> 1 + N (r 1)
> 1+
> .

This shows that for any given positive number we can find a term in the
sequence {rn } n
n=1 which is greater than the number. This means that r
as n .
If r < 1 then rn = (1)n (r)n with r > 1. Thus, as n becomes large, rn
alternates between large positive numbers and negative numbers with large
absolute value so that again the limit rn as n does not exist.
(c) If 0 < r < 1 then rn = (r1 1
)n
with (r1 )n as n . (See (b)).
Hence, rn 0 as n . If 1 < r < 0 then 0 < r < 1. In this case,
rn = (1)n (r)n 0 as n . If r = 0 then rn = 0 and limn rn = 0

Exercise 26.6
The series n1
P
n=1 ar is called a geometric series with ratio r.
(a) Show that
n
Sn = a 1r
1r
for r 6= 1.
Hint: Calculate Sn rSn .
a
(b) Show that the series converges to 1r
for |r| < 1 and diverges for |r| 1.

162
Solution.
n n
(a) We have Sn rSn = a ar so that Sn = a 1r 1r
for r 6= 1.
a
(b) If |r| < 1, using Exercise 26.5, we find limn Sn = 1r . If r = 1 then
Sn = na and this series diverges to either (if a > 0) or (if a < 0.)
If r = 1 the sequence {(1)n } n=1 is divergent and therefore the sequence

{Sn }n=1 is divergent.
The same applies if |r| > 1

Exercise 26.7 (Harmonic Series)

Consider the Harmonic series 1
P
n=1 n .
m
(a) Let n = 2 where m is a positive integer. Then
1 1 1
Sn =1 + + + + m
2  3 2  
1 1 1 1 1 1 1
=1 + + + + + + +
2 3 4 5 6 7 8
 
1 1
+ + + + m
2m1 + 1 2
Show that Sn 1 + m2 .
(b) Use (a) to show that limn Sn = . Thus, the Harmonic series is
divergent.

Solution.
(a) We have Sn 1+ 21 +2 14 +4 81 + +2m1 21m = 1+ 12 + 12 + + 21 =
1+ m2 .
(b) If n goes to then m goes to and therefore limm 1 + m2 = .
By (a), we conclude that limn Sn = and therefore the Harmonic series
is divergent

Exercise 26.8
Show that if
P P P
n=1 an = L1 and n=1 bn = L2 then n=1 (an + bn ) =
L1 + L2 for all , R.

Solution.
Let Sn2 = a1 +a2 + +an and Sn2 = b1 +b2 + +bn . Then lim n (Sn1 +
Sn2 ) = L1 + L2

Exercise 26.9  
Find the value of the infinite sum 3 5
P
n=1 4n + n(n+1)
.

163
Solution.P P 1
n1
We have 3 3 3 1
P 5
n=1 4n = 4 = 1 14
= 1 and = 5.
P  3 n=1 4 4 n=1 n(n+1)
5
Thus, n=1 4n + n(n+1) =6

Exercise 26.10
Show that the sequence { n2 1 n}
n=1 is convergent and find its limit.

Solution.
We have

( n 2 1 n)( n2 1 + n)
lim ( n2 1 n) = lim
n n ( n2 1 + n)
1
= lim =0
n 2
n 1+n
Exercise
P 26.11
Let n=1 an be a conditionally convergent series. bn = 12 (an +|an |) and
P DefineP
cn = 2 (an |an |). Prove that the two series n=1 and
1
n=1 are divergent.

Solution. P
P that |an | = bn cnPfor
Note first

all n N. Since P n=1 |an | is divergent,

either
P n=1 n b is divergent or n=1 |a n |. Suppose that n=1 bn is convergent
and n=1 cn is divergent. Since bn = |an | + cn we must have that
P
n=1 bn
is
Pdivergent which is a contradiction. A similar contradiction is obtained
P if
P
n=1
Pb n is divergent and c
n=1 n is convergent. It follows that both n=1 bn
and n=1 cn must be divergent

Exercise 26.12
Let Sn be the n-th partial sum of the series n2
P
n=1 n(n+1)(n+2) .
3 2 2
(a) Show that Sn = n+1 n+1 n+2 . Hint: Partial fractions.
P n2
(b) Find the value of the series n=1 n(n+1)(n+2) .

Solution.
(a) Using partial fraction decomposition we find that
n2 3 2 1
= .
n(n + 1)(n + 2) n+1 n+2 n

164
Thus,
n n n
3 X1 X1 X1 1 3 2 2
Sn = 3 2 1
2 i=3
i i=3
i i=3
i 2n+1 n+1 n+2
3 2 2
=
n+1 n+1 n+2
Thus, the series is convergent with sum

X n2
= lim Sn = 0
n=1
n(n + 1)(n + 2) n

Exercise 26.13 P
Let {an }
n=1 be a decreasing sequence such that n=1 an is convergent.
(a) Show that an 0 for all n N.
(b) Let  > 0. Show that there is a positive integer N such that if n > m N
we have
|am+1 + am+2 + + an | < .
(c) Show that (n N )an < .
(d) Let n > 2N. Show that n2 < n N.
(e) Show that na2n < .
(f) Show that limn nan = 0.

Solution.
(a) If aN < 0 for some N then an aN < P 0 for all n N. But then
limn an 6= 0 which contradicts the fact that n=1 an is convergent.
(b) This follows from the fact that the sequence of partial sums is Cauchy.
(c) From (b), we have aN +1 +aN +2 + +an < . But an aN +2 aN +1 .
Thus, (n N )an < .
(d) We have n > 2N n2 < N n2 < n N.
(e) If n > 2N > N we have na2n < (n N )an < .
(f) It follows from (e) that limn na2n = 0 and therefore limn nan =
2 limn na2n = 0

Exercise 26.14
Let N be P P Suppose that an = bn for all n N. Then
a positive integer.
the series a
n=1 n and n=1 bn either both converge or both diverge. Thus,
changing a finite number of terms in a series does not change whether or not
it converges, although it may change the value of its sum if it does converge.

165
solution.
The proof follows from the equality
P PN 1 P
n=1 an = n=1 (an bn ) + n=1 bn .

166
Solutions to Section 27
Exercise 27.1 (Comparison test))
Let {an }
n=1 and {bn }n=1 be two series such that 0 an bn for all n 1.
Let {Sn }n=1 be the sequence of partial sums of {an }

n=1 and {Tn }n=1 that of
{bn }
n=1 .
(a) Show that the sequences {Sn }
n=1 and {Tn }n=1 are increasing.
(b) Show that Sn Tn for all n 1.
(c) Show that if {bn }
n=1 is convergent then {Sn }n=1 and {Tn }n=1 are bounded.

(d) Show that if {bn }n=1 is convergent then {an }n=1 is also convergent.
(e) Show that if {an }
n=1 is divergent then {bn }n=1 is also divergent.

Solution.
(a) Sn+1 Sn = an+1 0 so that Sn+1 Sn . Thus, {Sn } n=1 is increasing.

Likewise, {Tn }n=1 is increasing.
(b) Since an bn for all n 1, we must have Sn Tn for all n 1.
(c) If {bn }
n=1 is convergent then the sequence {Tn }n=1 is also convergent. By

Exercise 3.9, the sequence {Tn }n=1 is bounded say by M. Since Sn Tn M
for all n 1, the sequence {Sn }n=1 is also bounded.

(d) Since {Sn }n=1 is increasing and bounded from above, it must be conver-
gent according to Exercise 5.5.
(e) This is just the contrapositive of (d)

Exercise 27.2
1 1
(a) Show that for n 1 we have (n+1)2 n(n+1) .
P 1
(b) Show that the series n=1 (n+1)2 is convergent.

Solution.
1
(a) Since 1 0 we have n + 1 n n+1 n1 (n+1)
1 1
2 n(n+1) .
P 1
(b) Since the series n=1 n(n+1) is convergent so does the given series

Exercise 27.3
Show that the series 1
P
n=1

n2 n+1
is divergent.

Solution.
Indeed, for n 1 we have n2 + (1 n) n2 so that n2 P n + 1 n2 = n.
This implies that 0 n1 n2 n+1
1
. Since the series 1
n=1 n is divergent
P 1
(harmonic series), the comparison test asserts that the series n=1 n2 n+1
is divergent.

167
Exercise
P 27.4 (Limit
P Comparison Test)
Let n=1 an and n=1 bn be two series with positive terms. Suppose that
an
lim =L>0
n bn

(a) Let  = L2 . Show that there exists a positive integer N such that
 
 an  L
 bn L  < 2 for all n N.

(b) Use (a) to establish

L
< an < 32 Lbn for all n N.
b
2 n

(c) Show that

P P
n=1 an is divergent if and only if n=1 bn is divergent.

Solution.
(a) This follows from the definition of convergence of a sequence.
(b) The inequality in (a) is equivalent to

L an L
L < <L+
2 bn 2
or
L an 3
< < L.
2 bn 2
or
L 3
bn < an < Lbn , n N.
2 2
P
(c)
PIf the series n=1 an converges then by the comparison
P test the series
L
n=1 2 bn is convergent. By Exercise 26.8, the series n=1 bn is also conver-
gent.
Conversely, if
P P 3
n=1 b n is convergent then n=1 2 Lbn is
Pconvergent and by
P
the comparison Ptest n=1 an is convergent. Similarly, n=1 an is divergent
if and only if n=1 bn is divergent.

Exercise 27.5
Determine whether the series 3n+1
P
n=1 4n3 +n2 2
converges or diverges.

168
Solution.
3n+1 3n 3 1
For large n we have an = 4n3 +n2 2
4n3
= 4n2
. So let bn = n2
. Then

an 3n3 + n2 3
lim = lim = .
n bn n 4n3 + n2 2 4
P 1
P 3n+1
Since the series n=1 n2 is convergent then so does the series n=1 4n3 +n2 2 .

Exercise 27.6

Let {aP
n }n=1 be a bounded sequence of nonnegative terms. Show that if the
series n=1 an is divergent so does the series
an
P
n=1 1+an . Hint: Comparison
test.

Solution.
Since the sequence {an }
n=1 is bounded there is M > 0 such that an M for
all n N. Thus,
an 1
an .
1 + an 1+M
But the series
P 1
P
n=1 an is divergent so does the series 1+M n=1 an . Hence,
by the comparison test, the series an
P
n=1 1+an is divergent

Exercise 27.7
Use the limit comparison test to show that the series 1
P
i=1 2n+ln n
is divergent.

Solution.
We have
1
2n+ln n n 1 1
lim 1 = lim = lim 1 = <1
n n 2n + ln n n 2 + 2
n n

Since the Harmonic series is divergent, the given series is also divergent

Exercise 27.8
Suppose that an 0 and that the series
P
n=1 an diverges. Suppose that
an
{an }n=1 is unbounded. Show that limn 1+an 6= 0. Hint:
P assume the con-
trary and get a contradction. Conclude that the series an
n=1 1+an is diver-
gent.

169
Solution.
an
If limn 1+a n
= 0 then we can find a positive integer N such that for
an
n N we have 1+a n
< 12 which implies that an < 1 for all n N. Let
M = a1 + a2 + + aN 1 + 1. Then |an | M for all n N. This shows that
the sequence is bounded contradicting the assumption that it is unbounded.
an
Thus, limn 1+a n
6= 0 and the given series is divergent

Exercise 27.9
Suppose that an 0 for all n N and that the series
P
n=1 an converges.
(a) Show that there is a positive
P 2 integer N such that a n < 1 for all n N.
(b) Show that the series n=1 an converges.

Solution.
(a) Since the series is convergent, we have limn an = 0. Thus, we can find
a positive integer N such that an < 1 for all n N.
(b) For n N we have a2n < an . By the comparison test the series 2
P
n=1 an
converges

Exercise 27.10
Use comparison test to show that the series
P
n=1 ( n2 + 1 n) is divergent.

Solution.
1
We have n2 + 1 n = n2 +1+n > 1+12 n1 . The Harmonic series is divergent
so that by the comparison test the given series is divergent

170
Solutions to Section 28
Exercise 28.1 (Alternating Series Test)
Let {an }n=1 be a sequence of positive numbers such that
(i) an an+1 , that is the sequence {an }n=1 is decreasing.
(ii) limn an = 0.
Let {Sn } sequence of partial sums of the series n1
P
n=1 bePthe n=1 (1) an .
n k1
That is, Sn = k=1 (1) ak .
(a) Show that for each n 1 we have S2n S2n+2 . That is, the sequence
{S2n }n=1 is increasing. Hint: Show that S2n+2 S2n 0.
(b) Show that the sequence {S2n+1 } n=1 is decreasing.
(c) Show that for all n 1, we have S2n a1 . Hence, the sequence {S2n } n=1
is bounded from above. Conclude that the sequence {S2n } n=1 is convergent,
say to L1 .
(d) Show that for all n 1, we have S2n+1 (a1 a2 ). Hence, the sequence
{S2n+1 }n=1 is bounded from below. Conclude that the sequence {S2n+1 }n=1

is convergent, say to L2 .
(e) Show that L1 = L2 . Hint: S2n+1 = S2n + a2n+1 .
(f) LetPL = L1 = L2 . Show that limn Sn = L. We conclude that the
series n=1 (1)
n1
an is convergent. Hint: Look at the sequence {cn }n=1 in
Exercise 10.4.
Solution.
(a) We have S2n+2 S2n = a2n+1 a2n+2 0. Thus, the sequence {S2n } n=1
is increasing.
(b) We have S2n+1 S2n1 = a2n+1 a2n 0. Thus, the sequence {S2n+1 }
n=1
is decreasing.
(c) We have
S2n = a1 (a2 a3 ) (a4 a5 ) (a2n2 a2n1 ) a2n a1 .
By Exercise 5.5, the sequence {S2n }
n=1 is convergent, say to L1 .
(d) We have
S2n+1 = (a1 a2 ) + (a3 a4 ) + + (a2n1 a2n ) + a2n+1 (a1 a2 ).
By Exercise ??, the sequence {S2n+1 } n=1 is convergent, say to L2 .
(e) Since S2n+1 = S2n + a2n+1 , we have L2 = limn S2n+1 = limn S2n +
limn a2n+1 = L1 + 0 = L1 .
(f) Since {Sn }
n=1 = {S1 , S2 , S3 , } with limn S2n = limn S2n+1 = L,
by Exercise 10.4(c), limn Sn = L

171
Exercise 28.2
(1)n1
Show that the series
P
n=1 n
is convergent.

Solution.
To see this, let an = n1 . Then n < n + 1 implies that n+11
< n1 that is
an+1 < an . Also, limn an = limn n1 = 0. Hence, by the previous result
the given series is convergent.

Exercise 28.3
(1)n1
Show that the series
P
n=1 n(n+1)
is convergent.

Solution.
1 1
To see this, let an = n(n+1) . Now n < n + 2 implies that n+2 < n1 and
1 1
this implies that (n+1)(n+2) < n(n+1) that is an+1 < an . Also, limn an =
1
limn n(n+1) = 0. Hence, by the alternating series test the given series is
convergent.

Exercise 28.4
Determine whether the series n1 n
P
n=1 (1) n+1
converges or diverges.

Solution.
Since
n
lim (1)n1 6= 0
n n+1
the series is divergent

Exercise 28.5
Determine whether the series n1 ln (4n)
P
n=1 (1) n
converges or diverges.

Solution.
Let f (x) = ln (4x)
x
for x 1. Then f 0 (x) = 1lnx2(4x) < 0 for x 1. This
shows that the sequence { ln (4n)
n
}
n=1 is decreasing. Moreover, we have by
using LHopitals rule

ln (4n) 1
lim = lim = 0.
n n n n

Hence, the given series is convergent

172
Exercise 28.6
n
(a) Show that nn! 1 for all
P n 1. n1 nn
(b) Show that ther series n=1 (1) n!
is divergent.

Solution.
(a) We have
nn n n n
= > 1.
n! nn1 1
n
(b) Since limn (1)n1 nn! 6= 0, the given series is divergent

Exercise 28.7
Show that the series n1 3n+1 +2n+1
P
n=1 (1) 3n n
diverges.

Solution.
We have

n

3n+1 + 2n+1 3n 3 + 2 32
lim = lim n

n 3n n n 3n 1 3n

n
3 + 2 23
= lim n
n 1 3n
3+0
= =3
10
n+1 +2n+1
Since limn (1)n1 3 3n n
6= 0 the given series is divergent

173
Solutions to Section 29
Exercise 29.1
(1)n1
Show that the series
P
n=1 n(n+1)
is absolutely convergent.

Solution.
This follows from Exercise 26.1

Exercise 29.2
Let
P P
n=1 n be an absolutely convergent series. Define the sequence
a P n=1 bn
by bn = |an | and note that an bn . Show that the sequence n=1 an is
convergent. That is, absolute convergence implies convergence.

Solution.
This is a direct consequence of the comparison test

Exercise 29.3
Give an example of a series that is convergent but not absolutely convergent.

Solution.
(1)n1
An example is the series
P
n=1 n(n+1)

Exercise 29.4
Give an example of a series that is conditionally convergent.

Solution.
(1)n1
An example is the series
P
n=1 n

Exercise 29.5P
Suppose that n=1 an is absolutely convergent.
(a) Show that 0 |an |+a
2
n
|an| and 0 |an |a
2
n
 |an | 
P
(b) Show that the series n=1 |an |+a |an |an
n
P
2
and n=1 2
are convergent.

Solution.
(a) Since |an | an we have |an |+a2
n
0. Since an |an |, we have an +|an |
|an |+an
2|an | or 2
|an |. Likewise, one can establish the second inequality.
P
(b) This follows from the comparison test and the fact that n=1 |an | is
convergent

174
Exercise 29.6 P
(a) Show that if
P 2
n=1 an is absolutely convergent then the series n=1 an is
also absolutely convergent.
(b) Give an example of a convergent series
P P 2
n=1 an for which n=1 an is
divergent.
Solution.
(a) Since limn |an | = 0 we can find a positive integer N such that for
n
P N we have |an | < 1.PThus, for n P N we have |an |2 < |an |.
PThus,
1 1 1
|an | = n=1 |an | + n=N |an | < n=1 |an | + n=N |an | < N
2 N 2 2 N 2 2 2
P P
Pn=1 n=1 |an | +

n=1 |an |. TheP series on the right is convergent so that by the comparison
test the series 2
n=1 |an | is convergent.
(1)n
(b) By the alternating series test the series
P
n=0 n+1 is convergent. But

the series 1
P
n=1 n is divergent

Exercise 29.7P

P n=1 an is absolutely convergent and {bn }n=1 is bounded.
Suppose that
Show that n=1 an bn is absolutely convergent (and thus convergent).
Solution.
|bn | M for all n N. Thus,P
P
LetPM > 0 such that n=1 |an bn |
M
P
n=1 |a n |. Since n=1 |a n | is convergent, the series
P M n=1 |an | is con-
vergentPand so by the comparison test the series n=1 |an bn | is convergent.
Thus, n=1 an bn is absolutely convergent and thus convergent

Exercise 29.8
Test the following series for absolute convergence, conditional convergence,
or divergence.
(a) P sin n
P
n=1 n2n .

(b) n=1 (1)n n25n +2n
.
n 2n 2n
P
(c) n=1 (1) 2n +2n .
Solution. 
n21n 21n . The geometric series

(a) We have  sin n 1
P
n2 n n=1 2n converges so
that by the comparison test the given series converges absolutely.
(b) By the limit comparison test one can show that the series of absolute value
is divergent. By the alternating series test, the given series is convergent.
Hence, the given series is conditionally convergent.
(c) The limit of the nth term is either 1 or 1 so that by the nth term test
the series is divergent

175
Exercise 29.9
Show that the series n1 ln 4n
P
n=1 (1) n
is absolutely convergent.

Solution.
By Exercise ?? we know that the series is convergent. However, since lnn4n >
1
n
and the Harmonic series is divergent, the series of absolute values is diver-
gent. Hence, the given series is conditionally convergent

Exercise 29.10
Suppose that the sequence {an }n=1 is monotone decreasing with limn an =
0. LetP{bn }
n=1 be a sequence such that |bn | an an+1 for all n N. Show

that n=1 bn is absolutely convergent.

Solution.
first that the conditions on an imply that an 0 for all n N. Let
Note P
Sn = nn=1 |bn |. Since |bn | an an+1 for each n we have

Sn (a1 a2 ) + (a2 a3 ) + + (an an+1 ) = a1 an+1 a1 .

It follows that the sequence {Sn }

n=1 is bounded from above by a1P
. Moreover,
this sequence is increasing.
P Hence, by Exercise 5.5, the series n=1 |bn | is
convergent so that n=1 bn converges absolutely

176
Solutions to Section 30
Exercise
P 30.1 (Integral Test)
Let n=1 an be a series of positive terms and suppose that there is a function
f : [1, ) R such that f is decreasing and positive with f (n) = an for all
n 1.
(a) Show that {Sn } n=1 is increasing. R
x
(b) Define F : [1, ) R by F (x) = 1 f (t)dt. Show that F is increasing.
R(c) For n 2 and x [n 1, n], show that an f (x) an1 and an
n
n1
f (x)dx an1 .
(d) Show that SnR a1 F (n) Sn1 .

(e) Suppose that 1 f (x)dx = L. Since F is increasing we can write F (n) L
for all n 1. Show that {Sn } n=1 is bounded. P
Hint: Use (d).

(f) Show that {Sn }n=1 is convergent. Hence, P n=1 an is convergent.
(g) Conversely, suppose that the series n=1 an converges to a number S.
Show that for any positive integer n 2 we have
F (n) S.
R
(h) Show that for all R 1 we have F (R) S. Thus, 1 f (x)dx =
RR
limR 1 f (x)dx is convergent. Hint: For any R 1 we have R [R] + 1
with [R] + 1 2.

Solution.
(a) We have Sn+1 Sn = an+1 > 0 so that Sn+1 > Sn .
(b) This follows from F 0 (x) = f (x) > 0.
(c) We have an = f (n) f (x) f (n 1) = an1 . Integrate from n 1 to n
we obtain the desired result. R2 R3
R n We have SnR
(d)
n
a1 = a2 + a3 + cdots + an 1 f (x)dx + 2 f (x)dx + +
n1
f (x)dx = 1 f (x)dx = F (n) a1 + a2 + + an1 = Sn1 .
(e) Using (d) we have Sn a1 F (n) L so that Sn L + a1 for all n 1.
That is, {Sn }n=1 is bounded.
(f) This follows from Exercise 5.5.
(g) From (d), we have F (n) Sn1 S.
R [R]+1
(h) We have F (R) F ([R] +R 1) = 1 f (x)dx S for
R all R 1. Hence,

letting R we find that 1 f (x)dx S. That is, 1 f (x)dx is conver-
gent. RR
Note that F is increasing and bounded from above so that limR 1 f (x)dx
exists

177
Exercise 30.2 (p-series)P
(a) Show that the series P 1
n=1 np is convergent for p > 1.
(b) Show that the series 1
n=1 np is divergent for p 1.

Solution.
If p < 0 then limn n1p = so that by the nth term test the series is
divergent. If p = 0 then the series is an infinite sum of 1 and so is divergent.
p
So suppose that p > 1. Let f (x) = x1p . Then f 0 (x) = xp+1 < 0 so that f
is decreasing on [1, )P and positive there. Moreover, fR(n) = n1p . So by the
dx
integral test the series 1
n=1 np converges if and only if 1 xp is convergent.
Now, we have Z R
dx R1p 1
p
= .
1 x 1p 1p
The improper integral exists if and only if p > 1. It is divergent if 0 < p <
1. Finally, for p = 1 the given series is just the Harmonic series which is
divergent

Exercise 30.3
Show that the series n
P
n=1 (n2 +1)(ln (n2 +1))a
is convergent for all a > 1. Hint:
The integral test.

Solution.
It is easy to check that the integrand is nonnegative and decreasing. By the
integral test we have
Z Z R
x x
2 2 a
dx = lim 2 2 a
dx
1 (x + 1)(ln (x + 1)) R 1 (x + 1)(ln (x + 1))
1 R
(ln (x2 + 1))1a 1

= lim
R 2(1 a)
 
1 1 1
= lim
R 2(1 a) (ln (R2 + 1)a1 (ln 2)a1
1
=
2(1 a)(ln 2)a1

Exercise 30.4
Use the integral test to test the convergence of the series 1
P
n=4 n ln n ln (ln n)
.

178
Solution.
We have Z Z
dx du
= = .
4 x ln x ln ln x ln ln 4 u
Thus, the given series is divergent

Exercise 30.5
Use the Integral Test to show that 2 n3
P
n=1 n e is convergent.

Solution.
Note that f (x) > 0 and f 0 (x) < 0 so that f is decreasing. By the Integral
Test we have
Z
3 1 3 1 1
x2 ex dx = lim [ eR + e1 ] = e1 .
1 R 3 3 3
Since the improper integral is convergent so does the given series

Exercise 30.6
Use the integral test to show that the series n2
P
n=1 e is convergent.

Solution.
2 2
Let f (x) = ex > 0. Then f 0 (x) = R2xex < 0 for x 1. Now, for
2
x 1 x2 x ex ex . Since 1 ex dx = 1e , the improper integral
R x2
1
e dx is convergent. Hence, by the integral test, the given series is
convergent

Exercise 30.7
(ln n)2
Use the integral test to show that the series
P
n=1 n
is divergent.

Solution. 2
Let f (x) = (lnxx) 0 for x 1. Moreover, f 0 (x) = 2 ln x ln x)2 x2 < 0 for
x 8. Thus, Z Z
(ln x)2 (ln x)2
.
1 x 8 x
But Z
(ln x)2 (ln R)3
= lim [ ] = .
8 x R 3
R 2
Since the improper integral 1 (lnxx) is divergent, the given integral is diver-
gent as well

179
Solutions to Section 31
Exercise 31.1 (Ratio Test)  
P  an+1 
Let n=1 an be a series of non-zero terms and suppose that limn  an  =
L 0.
(a) Suppose 0 L < 1. Let  = 1L
2
. Show that there is a positive integer N
such that
 
 an+1  1+L
 an  < 2 for all n N.

Hint: Use definition of convergence and Exercise 1.18.

(b) Let r = 1+L2
. Show that 0 < r < 1 and |aN +k | < rk |aN | for all k =
1, 2, .
(c) Find the value of the sum

X
rn |aN |.
n=1
Pn
(d) Let bn = k=1 |ak |. Show that the sequence {bn }
n=1 is increasing.
r|aN |
(e) Let M = bN + 1r . Show that |bn | M for all n 1.
(f) Show that the sequence {bn }
n=1 is convergent. What can you say about

the sequence {an }n=1 ?

Solution.  
 an+1 
(a) Since limn  an  = L, we can find a positive integer N such that
   
 an+1   an+1  1L
 an  L  an L  < 2 for all n N.

This is equivalent to
 
 an+1  L+1
 an  < 2
for all n N.

(b) Since L < 1 we have 1 + L < 2 and therefore r = 1+L 2

< 1. Clearly,
r > 0. By (a) we have aN +1 < r|aN |. Suppose that aN +k < rk |aN . Then
|aN +k+1 | < r|aN +K | < rk+1 |aN . Hence, |aN +k | < rk |aN for all k = 1, 2, .
(c) The given sum is a geometric series with ration r and whose sum is given
by

X X r|aN
rn |aN | = |aN | rn = .
n=1 n=1
1 r

180
(d) For any n 1, we have bn+1 = bn + |an+1 | bn . This shows that the
sequence {bn }n=1 is increasing.
(e) If n N then |bn | = nk=1 |ak | < M. If n N then |bn | = N
P P
Pn PN P PN k=1 |ak | +
r|aN
k=N |ak | k=1 |ak | + n=1 |an | = k=1 |ak | + 1r = M. It follows that

the sequence {bn }n=1 is convergent by Exercise ??. Now, by Exercise 29.2
the sequence {an }n=1 is convergent

Exercise 31.2 (Ratio Test)  

Let
P  an+1 
n=1 a n be a series of non-zero terms and suppose that limn  an 
=
L 0.
(a) Suppose L > 1. Let  = L 1. Show that there is a positive integer N
such that
 
L  an+1 <  for all n N.
 
an 

(b) Show that |an+1 | > |aP

N | > 0 for all n N.
(c) Show that the series n=1 an is divergent. Hint: The nth term test.

Solution.  
 an+1 
(a) Since limn  an  = L, we can find a positive integer N such that
 
L  an+1 < L 1 for all n N.
 
an 
 
 an+1 
(b) From (a) we conclude that  an  > 1 for all n N. Hence, |an+1 | >
|aN | > 0 for all n N.
(c) From P (b) we see that limn 6= 0 so that by the nth term test the
sequence n=1 an is divergent

Exercise 31.3
Consider the harmoninc series 1
P
n=1 n
which we know it is divergent. Find
limn an+1
an
.

Solution. 1
n+1
We have limn 1 =1
n

Exercise 31.4
Consider the series 1
P
n=1 n2 .
(a) Show that this series is convergent.
(b) Find limn an+1
an
.

181
Solution.
(a) This is a pseries with p = 2 > 1 so it is convergent.
1
(n+1)2
(b) We have limn 1 =1
n2

Exercise 31.5
Use the ratio test to determine the convergence of the series n 100n
P
n=1 (1) n!
.

Solution.
100n+1
(n+1)! 100
We have limn 100n = limn n+1
= 0 < 1. By the ratio test the series
n!
is convergent

Exercise 31.6 P 2n n!
Use the ratio test

n to determine the convergence of the series n=1 nn
. Hint:
limn 1 + n1 = e.

Solution.
2n+1 (n+1)!
(n+1)n+1 n

n 2
We have limn 2n n! = limn 2 n+1
= e
< 1. By the ratio test the
nn
series is convergent

Exercise 31.7
Find limn nn!2 .

Solution.
2
Let an = nn! . By the ratio test we have
an+1 n+1
lim = lim =0<1
n an n n2
P
so that the series n=1 an is convergent. Hence,
lim an = 0.
n

We conclude that
n! 1
lim2
= lim =
n n n an

Exercise 31.8 (nth root test)

1
Consider a series
P
n=1 an . Define L = limn |an | .
n

(a) Suppose first that 0 L < 1. Let  = 1L

2
. Show that there is a positive
integer N such that

182
 1 1+L
|an | n < 2
for all n N.

(b) Let r = 1+L

2
. Show that 0 <Pr < 1 and |an | < rn for all n N.

(c) Use (b) to conclude that n=1 an is absolutely convergent and hence
convergent.

Solution.
1
(a) Since limn |an | n = L, we can find a positive integer N such that
 
1 1
|an | L |an | L < 1L for all n N.
n
 n

2

This is equivalent to
1 1+L
|an | n < 2
for all n N.

(b) Since 0 L < 1 1 + L < 2 r = 1+L 2

< 1. Moreover,P by (a),

we have |an | < rn for all n N. (c) Since the goemetric series
P n=1 r
n

is convergent for 0 < r < 1, by the comparison test the series n=1 an is
absolutely convergent and hence convergent

Exercise 31.9 (nth root test)

Suppose that L > 1 in the previous exercise. Prove that the series
P
n=1 an
is divergent. Hint: nth term test.

Solution.
1
Let  = L1. Then there is a positive integer N such that ||an | n L| < L1.
1 1
This implies that L |an | n < L 1 or |an | n > 1. Hence, for n N we have
|an | > 1. But this means that limn 6= 0. Hence, by the nth term test, the
given series is divergent

Exercise 31.10
The root test is inconclusive ifPL = 1.
(a) We know that the series 1
n=1 n2 is absolutely convergent. Show that
L = 1.
(b) We know that the series n1 1
P
n=1 (1) n
is conditionally convergent. Show
that L = 1.
(c) We know that the series 1
P
n=1 n is divergent. Show that L = 1.

183
Solution.
(a) The given series is a pseries with p = 2 so it is convergent by Exercise

1
??. Let y = n12 n . Then ln y = 2 lnn n 0 as n . Thus, y 1 as
n . That is, L = 1.
(b) The given series is conditionally convergent by Exercise ??. Let y =

1
1 n
n
. Then ln y = lnnn 0 as n . Thus, y 1 as n . That is,
L = 1.
(c) The given series is the harmonic series so it is divergent. As in (b),
L=1

Exercise 31.11
Use the root test to show that the series nn
P
n=1 31+2n
is divergent.

Solution.
We have  n 1
 n n 1 n
lim  1+2n  = lim = .
n 3 3 n 9
By the root test the given series is divergent

Exercise 31.12  n
Use the root test to show that the series 5n3n3
P
n=1 7n3 +2
is absolutely con-
vergent.

Solution.
We have
 5n 3n3  3
 
1
lim |an | = lim  3
n  = < 1.
n n 7n + 2  7
By the root test the given series is convergent

184
Solutions to Section 32
Exercise 32.1
nx
Define fn : [0, ) R by fn (x) = 1+n 2 x2 . Show that the sequence {fn }n=1

converges pointwise to the function f (x) = 0 for all x 0.

Solution.
For all x 0, limn fn (x) = 0

Exercise 32.2
For each positive integer n let fn : (0, ) be given by fn (x) = nx.
Show that {fn }
n=1 does not converge pointwise on D.

Solution.
This follows from the fact that limn nx = for all x D

Exercise 32.3
For each positive integer n let fn : [0, 1] be given by fn (x) = nx . Show
that {fn }
n=1 converges uniformly to the zero function. Hint: For a given ,
choose N such that N > 1 .

Solution.
Let  > 0 be given. Let N be a positive integer such that N > 1 . Then for
n N we have
|x| 1 1
|fn (x) f (x)| = <
n n N
for all x [0, 1]

Exercise 32.4
nx
Define fn : [0, ) R by fn (x) = 1+n 2 x2 . By Exercise 32.1, this sequence

converges pointwise to f (x) = 0. Let  = 13 . Show that there is no positive

integer N with the property n N implies |fn (x) f (x)| <  for all x 0.
Hence, the given sequence does not converge uniformly to f (x).

Solution.
For any positive integer N and for n N we have
    
fn 1 f 1  = 1 > 
 
 n n  2

185
Exercise 32.5
Define fn : [0, 1] R by fn (x) = xn . Define f : [0, 1] R by

0 if 0 x < 1
f (x) =
1 if x = 1

(a) Show that the sequence {fn }

n=1 converges pointwise to f.
(b) Show that the sequence {fn }
n=1 does not converge uniformly to f. Hint:
Suppose otherwise. Let  = 0.5 and get a contradiction by using a point
1
(0.5) N < x < 1.

Solution.
(a) For all 0 x < 1 we have limn fn (x) = limn xn = 0. Also,
limn fn (1) = 1. Hence, the sequence {fn }
n=1 converges pointwise to f.
1
(b) Suppose the contrary. Let  = 2 . Then there exist a positive integer N
such that for all n N we have
1
|fn (x) f (x)| <
2
for all x [0, 1]. In particular, we have
1
|fN (x) f (x)| <
2
1
for all x [0, 1]. Choose (0.5) N < x < 1. Then |fN (x) f (x)| = xN > 0.5 =
 which is a contradiction. Hence, the given sequence does not converge
uniformly

Exercise 32.6
Give an example of a sequence of continuous functions {fn }
n=1 that converges
pointwise to a discontinuous function f.

Solution.
See Exercise 32.5(a)

Exercise 32.7
Suppose that for each n 1 the function fn : D R is continuous in D.
Suppose that {fn }
n=1 converges uniformly to f. Let a D.
(a) Let  > 0 be given. Show that there is a positive integer N such that if
n N then |fn (x) f (x)| < 3 for all x D.

186
(b) Show that there is a > 0 such that for all |x a| < we have |fN (x)
fN (a)| < 3
(c) Using (a) and (b) show that for |x a| < we have |f (x) f (a)| < .
Hence, f is continuous in D since a was arbitrary. Symbolically we write

lim lim fn (x) = lim lim fn (x)

xa n n xa

Solution.
(a) This follows from the definition of uniform convergence.
(b) This follows from the fact that fN is continuous at a D.
(c) For |x a| < we have |f (x) f (a)| = |f (a) fN (a) + fN (a) fN (x) +
fN (x)f (x)| |fN (a)f (a)|+|fN (a)fN (x)|+|fN (x)f (x)| < 3 + 3 + 3 =


Exercise 32.8
Consider the interval [0, 1] and let the rationals in this interval be labeled
r1 , r2 , arranged in increasing order. For each positive integer n we define
the function fn : [0, 1] R by

1 if x {r1 , r2 , , rn }
fn (x) =
0 otherwise.

(a) Show that fn is Riemann integrable on [0, 1]. Hint: Remark 3.

(b) Show that {fn }
n=1 converges pointwise to the function

1 if x is rational
f (x) =
0 if x is irrational

(c) Show that f is not Riemann integrable.

Solution.
(a) Since fn is bounded and discontinuous at a finite number of points, by
Remark 3, fn is Riemann integrable.
(b) If x is irrational then for every n 1 we have fn (x) = 0. Thus, limn fn (x) =
0 = f (x). Suppose now that x is rational. Then there is a positive in-
teger k such that x = rk . In this case, fn (x) = 1 for all n k. Hence,
limn fn (x) = 1 = f (x). It follows that the given sequence converges point-
wise to f.
(c) This follows from Exercise 20.5

187
 Z Z Z
lim fn (x)dx = lim fn (x)dx = f (x)dx? (3)
n D D n D
That is, can we interchange limit and integration? The answer is no as seen
in the next exercise.

Exercise 32.9
Consider the functions fn : [0, 1] defined by fn (x) = n2 xenx .
(a) Show that {fn }n=1 converges pointwise to f (x) = 0. Hint: LH opitals
rule. R1
(b) Find limn 0 fn (x)dx. Hint: Integration by parts.
R1 R1
(c) Show that limn 0 fn (x)dx 6= 0 limn fn (x)dx.

Solution.
(a) We have
n2 x
lim fn (x) = lim n2 xenx = lim =0
n n n enx

where we apply LHopitals rule twice. Hence, {fn }

n=1 converges pointwise
to f (x) = 0.
(b) Using integration by parts we find
Z 1 1
xenx

nx 1 nx
2
n xe dx = n 2
2e = 1 en (n + 1).
0 n n 0

Hence, Z 1
lim fn (x)dx = lim [1 en (n + 1)] = 1.
n 0 n

(c) From (b) we have

Z 1 Z 1
lim fn (x)dx = 1 6= 0 = lim fn (x)dx
n 0 0 n

Exercise 32.10
Let {fn }n=1 be a sequence of Riemann integrable functions on [a, b] that
converges uniformly to a f defined on [a, b].
(a) Let  > 0 be given. Show that there is a positive integer N such that for
all n N we have

|fn (x) f (x)| < 4(ba)
for all x [a, b].

188
(b) Let n N. Show that there is a partition P of [a, b] such that

U (fn , P ) L(fn , P ) < .
2
(c) Suppose n N and P as in (b). Show that

U (f, P ) U (fn , P ) +
4
and therefore

L(f, P ) L(fn , P ) .
4
 
Hint: |f (x)| |fn (x)| + 4(ba) and |fn (x)| |f (x)| + 4(ba) (d) Conclude that
U (f, P ) L(f, P ) <  and therefore f is Riemann integrable on [a, b].

Solution.
(a) This is just the definition of uniform convergence.
(b) This follows from Exercise 20.6.

(c) Since |f (x)| |fn (x)| + 4(ba) we have

 
U (f, P ) U (fn , P ) + U ( , P ) = U (fn , P ) + .
4(b a) 4

Likewise, since |fn (x)| |f (x)| + 2(ba)

 
L(f, P ) L(fn , P ) + L( , P ) = L(fn , P ) .
4(b a) 4

Hence,
  
U (f, P ) L(f, P ) [U (fn , P ) L(fn , P )] + < + = .
2 2 2
It follows that f is Riemann integrable on [a, b]

Exercise 32.11
Let {fn }
n=1 and f be as in the previous exercise.
(a) Let  > 0 be given. Show that there is a positive integer N such that if
n N then

|fn (x) f (x)| ba
for all x [a, b].

189
(b) Show that for every n N we have
Z b Z b 
 

 fn (x)dx f (x)dx  < .

a a

Thus, (3) holds. Hint: Exercise 23.1 and Exercise 24.3

Solution.
(a) This follows from the definition of uniform convergence.
(b) Using (a) and Exercise 24.3 we find
Z b Z b  Z b 
   

 fn (x)dx f (x)dx  =  [fn (x) f (x)]dx
  
a a a
Z b
|fn (x) f (x)|dx
a
Z b

< dx = 
ba a

Exercise 32.12
Give an example of a sequence of differentiable functions {fn }
n=1 that con-
verges pointwise to a non-differentiable function f.

Solution.
See Exercise 32.5(a)

Exercise 32.13 q
Consider the family of functions fn : [1, 1] given by fn (x) = x2 + n1 .
(a) Show that fn is differentiable for each n 1.
(b) Show that for all x [1, 1] we have
1
|fn (x) f (x)|
n
q
where f (x) = |x|. Hint: Note that x2 + n1 + x2 1n .
(c) Let  > 0 be given. Show that there is a positive integer N such that for
n N we have

|fn (x) f (x)| <  for all x [1, 1].

190
Thus, {fn }
n=1 converges uniformly to the non-differentiable function f (x) =
|x|.

Solution.
(a) fn is the composition of two differentiable functions so it is differentiable
with derivative   1
0 2 1 2
fn (x) = x x + .
n
(b) We have
r   q 2 1 2 q 2 1 2 

2
1 2   ( x + n x )( x + n + x ) 
|fn (x) f (x)| =  x + x  = 

n
q
x2 + n1 + x2
   

1
n
=q
1
x2 + n
+ x2
1
n
notag (4)
1
n

(c) Let  > 0 be given. Since limn 1n = 0 we can find a positive integer
N such that for all n N we have 1n < . Now the answer to the question
follows from this and part (b)

Exercise 32.14
Give an example of a sequence of differentiable functions {fn } n=1 that con-
verges uniformly to a a differentiable function f such that limn fn0 (x) 6=
f 0 (x) = [limn fn (x)]0 . That is, one cannot, in general, interchange limits
and derivatives. Hint: Exercise 32.3

Solution.
Exercise ?? with limn fn0 (x) = g(x) where g(1) = 1 and g(x) = 0 for
0x<1

Exercise 32.15
Let {fn }
n=1 defined on a set D be uniformly Cauchy.
(a) Show that for each x D, the sequence of numbers {fn (x)} n=1 is con-
vergent. Call the limit f (x). Thus, we can define a function f : D R such

191
that f (x) = limn fn (x). Hint: Exercise 7.7
(b) Show that {fn }n=1 converges pointwise to f.
(c) Let  > 0 be given. Show that there is a positive integer N such that for
all m, n N we have

|fm (x) fn (x)| < 2
for all x D.

(d) Fix x D. Show that there is a positive integer m N such that

|fm (x) f (x)| < 2 .
(e) For the fixed x in (d), let n N. Show that |fn (x) f (x)| < .
(f) Conclude that {fn }n=1 converges uniformly to f.

Solution.
(a) Let x D. By uniform Cauchy, the sequence {fn (x)} n=1 is a Cauchy
sequence of numbers in R. By Exercise 7.7 it is convergent. Call its limit
f (x). Hence, we define a function f : D R by f (x) = limn fn (x).
(b) This follows from the definition of f.
(c) This follows from the definition of uniform Cauchy sequence.
(d) Fix x D. The result follows from the fact that {fn } n=1 converges
pointwise to f.
(e) For n N we have

|fn (x) f (x)| |fn (x) fm (x)| + |fm (x) f (x)| < .

(f) Since the previous inequality is true for all x D we conclude that
{fn }
n=1 converges uniformly to f

Exercise 32.16
Let {fn }n=1 be a sequence of differentiable functions on [a, b] such that
{fn (c)} 0
n=1 converges for some c [a, b]. Assume also that {fn }n=1 converges
uniformly to g in [a, b].
(a) Let  > 0 be given. Show that there is a positive integer N1 such that for
all m, n N1 we have
0
|fm (x) fn0 (x)| < 
2(ba)
for all x [a, b].

(b) Show that there is a positive integer N2 such that for all m, n N2 we
have

|fm (c) fn (c)| < .
2

192
Hint: Exercise 7.3
(c) Show that for all x [a, b] there is a d between c and x such that
0
fm (x) fn (x) = fm (c) fn (c) + (x c)[fm (d) fn (d)].

Hint: Apply the Mean Value theorem to the function fm fn restricted to

the interval [c, x].
(d) Let N = N1 + N2 . Use (a) - (c) to show that for n N we have

|fm (x) fn (x)| <  for all x [a, b].

That is, the sequence {fn }

n=1 is uniformly Cauchy.
(e) Show that the sequence {fn }n=1 converges uniformly to a a function f.

Solution.
(a) Let  > 0 be given. There is a positive integer N 0 such that for all n N1
we have

|fn0 (x) g(x)| < 

4(ba)
for all x [a, b].

Hence, for all m, n N1 we have

0   
|fm (x)fn0 (x)| |fm
0
(x)g(x)|+|fn0 (x)g(x)| < + =
4(b a) 4(b a) 2(b a)

(b) This follows from the fact that {fn (c)}

n=1 is Cauchy by Exercise 7.3.
(c) This is a result from the MVT.
0
(d) For n N we have |fm (x) fn (x)| |fm (c) fn (c)| + (b a)|fm (d)
0  
fn (d)| < 2 + (b a) 2(ba) = . Hence, the sequence {fn }n=1 is uniformly
Cauchy.
(e) This is a consequence of Exercise 32.15

Exercise 32.17
In this exercise we want to show that f of the previous exercise is differen-
tiable in [a, b] and f 0 = g.
(a) Show that there is a positive integer N1 such that for all n N1 we have
0
|fm (x) fn0 (x)| < 
3
for all x [a, b].

193
(b) Let x0 [a, b]. Use the MVT to the function fm fn to show the existence
of a point d between x0 and x such that
0
fm (x) fn (x) = fm (x) fn (x0 ) + (x x0 )[fm (d) fn0 (d)].

(c) Use (a) and (b) to show that

 
 fm (x) fn (x0 ) fn (x) fn (x0 )  
 < .
 x x0 x x0  3

(d) Show that  

 f (x) f (x0 ) fn (x) fn (x0 )  

 x x0  .
x x0  3
(e) Show that there is a positive integer N2 such that for all n N2 we have

|fn0 (x0 ) g(x0 )| < .
3
(f) Let N = N1 + N2 . Show that there is a > 0 such that
 
If 0 < |x x0 | < then  fN (x)f N (x0 ) 0  
f (x 0  < 3.
)

xx0 N

(g) Use (d) - (f) to conclude that

 
If 0 < |x x0 | < then  f (x)f (x0 )
g(x 0  < .
)
 
xx0

That is, f is differentiable at x0 with f 0 (x0 ) = g(x0 .

Solution.
(a) Similar to (a) in the previous exercise.
(b) Easy.
(c) We have
 
 fm (x) fn (x0 ) fn (x) fn (x0 )  0 
  = |fm (d) fn0 (d)| < .
 x x0 x x0  3

(d) This follows by letting m .

(e) This follows from the fact that {fn0 (c)}
n=1 converges to g(x0 ).

194
(f) This follows from the fact that fN is differentiable at x0 .
(g) Suppose 0 < |x x0 | < . Then
  
 f (x) f (x0 )   f (x) f (x0 ) fN (x) fN (x0 )

 x x0 g(x 0 ) =
  x x0
x x0

fN (x) fN (x0 ) 0 0

+ fN (x0 ) + fN (x0 ) g(x0 )
x x0
  
< + + =
3 3 3
Hence, f is differentiable at x0 with f 0 (x0 ) = g(x0 )
Exercise 32.18
n
Consider the sequence of functions fn (x) = x xn defined on [0, 1).
(a) Does {fn }n=1 converge to some limit function? If so, find the limit func-
tion and show whether the convergence is pointwise or uniform.
(b) Does {fn0 }
n=1 converge to some limit function? If so, find the limit func-
tion and show whether the convergence is pointwise or uniform.
Solution.
(a) Let  > 0 be given. Let N be a positive integer such that N > 1 . Then
for n N
n n
 
x x x = |x| < 1 1 < .
 
 n  n n N
Thus, the given sequence converges uniformly (and pointwise) to the function
f (x) = x.
(b) Since limn fn0 (x) = 1 for all x [0, 1), the sequence {fn0 }
n=1 converges
pointwise to f (x) = 1. However, the convergence is not uniform. To see
this, let  = 12 and suppose that the convergence is uniform. Then there is a
positive integer N such that for n N we have
1
|1 xn1 1| = |x|n1 < .
2
In particular, if we let n = N + 1 we must have xN < 12 for all x [0, 1).

1
But x = 21 N [0, 1) and xN = 12 which contradicts xN < 21 . Hence, the
convergence is not uniform
Exercise 32.19
Suppose that each fn is uniformly continuous on D and that fn f uni-
formly on D. Prove that f is uniformly continuous on D.

195
Solution.
Let  > 0 be given. By uniform convergence, there is a positive integer N
such that |fn (x) f (x)| < 3 for all x D whenever n N. Now, since fN is
uniformly continuous on D we can find a > 0 such that if |x y| < then
|fn (x) fN (y)| < 3 . Thus, for any x, y D such that |x y| < we have

|f (x) f (y)| =|f (x) fN (x) + fN (x) fN (y) + fN (y) f (y)|

|f (x) fN (x)| + |fN (x) fN (y)| + |fN (y) f (y)|
  
< + + =
3 3 3
Hence, f is uniformly continuous on D

Exercise 32.20
xn
Let fn (x) = 1+x n for x [0, 2].

(a) Find the pointwise limit f (x) = limn fn (x) on [0, 2].
(b) Does fn f uniformly on [0, 2]?

Solution.
(a) The pointwise limit is

0 if 0 x < 1
1
f (x) = if x = 1
2
1 if 1 < x 2
(b) The convergence cannot be uniform because if it were f would have to
be continuous

Exercise 32.21
Prove that if fn f and gn g uniformly on a set D then fn + gn f + g
uniformly on D.

Solution.
Let  > 0 be given. Let N1 be a positive integer such that |fn (x) f (x)| < 2
for all n N1 and all x D. Likewise, let N2 be a positive integer such that
|gn (x) g(x)| < 2 for all n N2 and all x D. Let N = N1 + N2 . Then for
n N we have
 
|(fn (x) + gn (x)) (f (x) + g(x))| < + = 
2 2
for all x D. That is, fn + gn f + g uniformly on D

196
Exercise 32.22
Prove that if fn f uniformly on a set D then {fn }
n=1 uniformly Cauchy
on D.

Solution.
Let  > 0 be given. Then there is a positive integer N such that if n N
then |fn (x) f (x)| < 2 for all x D. Thus, if m, n N we have
 
|fn (x) fm (x)| |fn (x) f (x)| + |fm (x) f (x)| < + =
2 2
for all x D. This shows that {fn }
n=1 uniformly Cauchy on D

Exercise 32.23
Suppose that {fn }
n=1 is uniformly convergent on a set D where each fn is
bounded on D, that is |fn (x)| Mn for all x D. Show that there is a
positive constant M such that |fn (x)| M for all n N and all x D.

Solution.
By the previous exercise {fn }
n=1 is uniformly Cauchy on D. Let  = 1. Then
there is a positive integer N such that

|fn (x) fN (x)| < 1 for all n N and all x D.

Hence,

|fn (x)| < 1 + MN for all n N and all x D.

Let M = 1 + M1 + M2 + + MN 1 . Then |fn (x)| M for all x D and

all n N

Exercise 32.24
Suppose that fn f and gn g uniformly on D. Moreover, suppose that
|fn (x)| Mn and |gn (x)| Mn for all n N and all x D. Prove that
fn gn f g uniformly on D.

Solution.
By the previous exercise there exist M1 > 0 and M2 > 0 such that |fn (x)|
M1 and |gn (x)| M2 for all x D and all n N. Let M = M1 + M2 .
Then |fn (x)| M and |gn (x)| M for all x D and all n N. Now,
|f (x)| |fn (x) f (x)| |f (x) fn (x) + f (x)| = |fn (x)| M |f (x)|

197
M + |fn (x) f (x)| for all n N and all x D. By pointwise convergence
we obtain |f (x)| M for all x D. Likewise, |g(x)| M for all x D.
Let  > 0 be given. Then there is a positive integer N1 such that if n N1

then |fn (x)f (x)| < 2M for all x D. Likewise, there is a positive integer N2

such that if n N2 then |gn (x) g(x)| < 2M for all x D. Let N = N1 + N2 .
Then for n N we have

|fn (x)gn (x) f (x)g(x)| |fn (x)||gn (x) g(x)| + |g(x)||fn (x) f (x)|
 
<M +M =
2M 2M
for all x D. This shows that fn gn f g uniformly on D

Exercise 32.25
2
Let fn (x) = x + n1 for all x R and gn (x) = x + n1 .
(a) Show that fn f uniformly where f (x) = x.
(b) Show that gn does not converge uniformly to the function g(x) = x2 .

Solution.
(a) Let  > 0 be given. Choose N such that N > 1 . Then for n N we have
1 1
|fn (x) f (x)| = < .
n N
This shows that fn converges uniformly to f.
(b) Suppose the contrary. Let  = 1. Then there positive integer N such that
if n N we have  
 2x 1 
|gn (x) g(x)| =  + 2  < 1
n n
for all n N and all x R. But if we choose n = N and x = N we obtain
2N 2 + 1
|gN (N ) g(N )| = >1
N2
a contradiction. Hence, gn does not convergen to g uniformly

Exercise 32.26
Give an example of a sequence {fn }
n=1 and a function f such that fn f
uniformly but fn does not converge uniformly to f 2 .
2

Solution.
1
Let fn (x) = x + n
and use previous exercise

198
Exercise 32.27
Give an example of two sequences {fn }
n=1 and {gn }n=1 such that fn f
and gn g uniformly but fn gn does not converge uniformly to f g. Thus,
the condition of boundedness in Exercise 32.24 is crucial.

Solution.
1
Let fn (x) = gn (x) = x + n
and f (x) = g(x) = x and use previous exercise

199
Solutions to Section 33
Exercise 33.1P
Suppose that n
n=0 an (x a) is a power series that converges for x = c.
Note that the series converges to a0 if c = a. So we will assume that c 6= a.
(a) What is the value of the limit limn an (c a)n ?
(b) Show that there is a positive integer N such that |an (c a)n | < 1 for all
n N. P 1
(c) Let M = N n n
n=0 |an (c a) | + 1. Show that |an (c a) | M for all n 0.
(d) Let x be such that |x a| < |c a|. Show that for any n 0 we have
 x a n
 
n
|an (x a) | M   .
c a
 xa n
(e) Show that the series P
P
n=0 M 
c
 is convergent.

(f) Show that the series n=0 an (x a)n is absolutely convergent and hence
convergent.
We conclude that if a power series n
P
n=0 an (x a) converges for x = c it is
convergent for any x satisfying |x a| < |c a|.

Solution.
(a) By the nth term test, limn an (c a)n = 0.
(b) From the definition of convergence of a sequence and part (a), there exists
a positive integer N such that |an (c P a)n | < 1 for all n N.
1
(c) If n N 1 then |an (c a)n | N n
n=0 |an (c a) | < M. If n N we
n
have |an (c a) | < 1 < M.
(d) For n 0 we have
 x a n  x a n
   
n n
|an (x a) | = |an (c a) | 
  M   .
c a c a
 
(e) This series is a geometric series with ratio  xa ca
 < 1 so it is convergent.
(f) This follows from the comparison test and Exercise 29.2

Exercise 33.2P
n
Suppose that n=0 an (x a) is a power series that diverges for x = d.
Let
P x be a number satisfying |x a| > |d a|. Show that the assumption
n
Pn=0 an (x a) converges at x leads to a contradiction. Hence, the series
n
n=0 an (x a) must be divergent. Hint: Use Exercise 33.1.

200
Solution.
|d a|. If n
P
Suppose that |x a| > P n=0 an (x a) is convergent then by
Exercise 33.1, the series n
n=0 an (d a) is absolutely
P convergent and hence
convergent.
P But this contradicts the fact that n=0 an (d a)n is divergent.
n
Hence, n=0 an (x a) must be divergent

Exercise 33.3
Consider a power series n
P
Pan (x a) . Let
n=0 C be the collection of all real
numbers at which the series n=0 an (x a)n converges. That is,

X
C = {x R : an (x a)n converges}.
n=0

(a) Show that C 6= .

(b) Explain in words the meaning that C = {a}.
(c) Explain in words the meaning that C = (, ) = R.
(d) SupposeP that C 6= {a} and C 6= R. That is, there is a real number d 6= a
such that a
n=0 n (da) n
diverges. Show that if x C then |xa| |da|.
Conclude that {|x a| : x C} is bounded from above with an upper bound
M. What is the value of M ?
(e) Show that there is a finite number R such that R is the least upper bound
of {|x a| : x C}. Thus, |x a| R for all x C. Show that R > 0.
PShow that nfor any real number x such that |x a| > R, the series
(f)
n=0 an (x a) is divergent.

PShow that nfor any real number x such that |x a| < R, the series
(g)
n=0 an (x a) is convergent. Hint: Let  = R |x a| and use the def-
inition of supremum to show that there exist an x0 C such that R  <
|x0 a| R.

Solution.
(a) Since the series converges at x = a to a0 , we have a C and so C 6= .
(b) If C = {a} the series converges only at x = a and diverges for all x 6= a.
(c) If C = (, ) then the series converges for all values of x.
(d) This follows from Exercise 33.2. Thus, for any x C we have |x a|
|d a|. This shows that the set {|x a| : x C} bounded from above by
M = |d a|.
(e) The existence of R follows from the completeness axiom of R. Since
|x a| R for all x C and a, d C with d 6= a we conclude that R > 0.

201
(f) Let x R such that |xa| > R. Then x 6 C and therefore n
P
n=0 an (xa)
is divergent.
(g) Let x R such that |x a| < R. Let  = R |x a|. By the definition of
supremum, there exists an x0 C such that R  < |x0 P a| R. But this
implies that |x a| < |x0 a|. By Exercise 33.1, the series n=0 an (x a)
n

is (absolutely) convergent

Exercise 33.4
FindPthe radius of convergence of each of the following series:
(a) P xn
n=0 n! .
(b) P
n=0 n!x .
n
n
(c) n=0 x .

Solution.
(a) By the ratio test the series converges for any value of x. Thus, R = .
(b) By the ratio test the series converges only when x = 0. Thus, R = 0.
(c) This is a geomteric series that converges for |x| < 1 so that R = 1

Exercise 33.5P
Suppose that n
n=0 an (x a) is a power series with an 6= 0 for all n 0.
Suppose that  
 an+1 
lim   = L 0.
n  an 
 
(xa)n+1 
(a) Find limn  an+1 .

an (xa)n 
(b) Suppose that L = 0. Show that R = . That is, a power series converges
for all x R.
(c) Suppose that L > 0. Show that R = L1 .
(d) Suppose that L = . Show that R = 0, that is, the series diverges for
all x 6= a.

Solution.
(a) By the ratio test we have
 an+1 (x a)n+1 
   
 an+1 
lim   = lim   |x a| = L|x a|.
n  an (x a)n  n  a)n 

(b) If L = 0 then L|x a| = 0 < 1 for all x R. Hence, R = and the

power series converges for all x R.

202
(c) A power series converges if and only if L|x a| < 1 and diverges for
L|x a| > 1. Thus, R = L1 .
(d) If L = > 1 then the series diverges for all x 6= a. That is, R = 0

Exercise 33.6 n
Find the interval of convergence of the power series n (x1)
P
n=1 (1) n
.

Solution.
We have R1 = limn n+1n
= 1 so that R = 1. Hence, the series converges for
|x 1| < 1 and diverges for |x 1| > 1. So the series converges for all x such
that 0 < x < 2. What about the Pendpoints x = 0 and x = 2? If we replace
1
x by 0 we obtain the series n=1 n which is divergent P (harmonic series).
If we replace x by 2 we obtain the alternating series n=1 (1) n1 1
n
which
converges by the alternating series test. Thus, the interval of convergence is
0<x2

Exercise 33.7
Find the interval of convergence of the power series n
xn .
P
n=1 n2 +2

Solution.
Let R be the radius of convergence. Then

1 n+1 n2 + 2
= lim = 1.
R n (n + 1)2 + 2 n

Thus, R = 1 so the series converges for |x| < 1 and diverges for |x| > 1. If
x = 1 then the series becomes n
P
n=1 n2 +2 . Since

n
n2 +2
lim 1 =1>0
n
n
P 1
and
P thenseries n=1 n is divergent, by the limit comparison test, the series
n=1 n2 +2 is also divergent.
(1)n n
If x = 1 then we get the alternating series
P
n=1 n2 +2 . By the alternating
series test one can show that this series is convergent. Thus, the interval of
convergence is 1 x < 1

Exercise 33.8
(1)n1 xn
Find the interval of convergence of the power series
P
n=1 n2 +1
.

203
Solution.
Use the absolute ratio test:
|x|n+1 n2 + 1 n2 + 1
   
 an+1 
 an  (n + 1)2 + 1 |x|n = |x| n2 + 2n + 2 |x| as n .
 =

Therefore the series converges absolutely if |x| < 1 and diverges if |x| > 1.
If |x| = 1, the ratio test gives no information, so we have to look at the
endpoints separately:
(1)n1
At x = 1 we have the series
P
n=1 n2 +1 an absolutely convergent series
by the alternating series test.
At x = 1 we have the series 1
P
an absolutely convergent series
P 1n=1 n2 +1
by comparison to the series n=1 n2 .
Conclusion: This power series is absolutely convergent on [1, 1] and diverges
everywhere else

Exercise 33.9
e n (x1)n
Find the interval of convergence of the power series
P n1


n=1 (1) 2 n
.

Solution.
Use the absolute ratio test:
 an+1  en+1 |x 1|n+1 2n n
 
e n e
 =
n+1
n n
= |x 1| |x 1| as n .
 an  2 (n + 1) e |x 1| 2 n+1 2

Therefore the series converges absolutely if |x 1| < 2e and diverges if

|x 1| > 2e . If |x 1| = 2e , the ratio test gives no information, so we have to
look at the endpoints separately:
At x 1 = 2e we have the series 1
P
n=1 n a divergent series.
(1)n1
At x 1 = 2e we have the series
P
n=1 n
a conditionally convergent
series.
Conclusion: This power series is absolutely convergent on |x 1| < 2e , con-
ditionally convergent at x = 2e and divergent everywhere else

Exercise 33.10
Suppose that the power series n
P
n=0 an x converges if x = 3 and diverges
if x = 7. Indicate which of the following statements must be true, cannot be
true, or may be true.
(a) The power series converges if x = 10.

204
(b) The power series diverges if x = 3.
(c) The power series converges if x = 6.
(d) The power series diverges if x = 2.
(e) The power series diverges if x = 7.
(f) The power series converges if x = 4.

Solution.
The thing we know for sure is that the series converges for |x| < 3 and
diverges for |x| > 7.
(a) Cannot be true.
(b) May be true.
(c) May be true.
(d) Cannot be true.
(e) May be true.
(f) May be true

Exercise 33.11
Give an example of a power series that converges on the interval [11, 3).

Solution.
(x+7)n
An example is the series
P
n=1 n4n

Exercise 33.12
Determine all the values of the real number x for which the series

X xn
n=1
3n n(log (3n))3
converges

Solution.
The radius of convergence is
   3
 an 
  = 3 n + 1 log (3n + 3) 3 as n .
 an+1  n log 3n
So the series converges for all x satisfyin |x| < 3. If x = 3 the series becomes
P (1)n
n=1 n(log 3n)3 which converges by the alernating series test. If x = 3 the
series becomes 1
P
n=1 n(log 3n)3 which converges by the integral test. Thus, the
given series converges on the interval [3, 3] and diverges elsewhere

205
Solutions to Section 34
Exercise 34.1
1
Find the Taylor series of f (x) = 1x
, where 1 < x < 1.

Solution.
Finding successive derivatives we obtain

f 0 (x) =(1 x)2 f 0 (0) = 1

f 00 (x) =2(1 x)3 f 00 (0) = 2 = 2!
f 000 (x) =3 2(1 x)4 f 000 (0) = 3!
..
.
f (n) (x) =n (n 1) 3 2(1 x)(n+1) f (n) (0) = n!

Thus,
1 xn+1
Pn (x) = 1 + x + x2 + x3 + + xn = .
1x
This is the nth partial sum of a geometric series that converges for |x| < 1.
Moreover,
1
1 + x + x2 + x3 + = .
1x
This shows that
1
= 1 + x + x2 +
1x
for all 1 < x < 1

Exercise 34.2
Consider the function

0 if x = 0
f (x) = 2
e x2 6 0
if x =

(a) Find the Taylor polynomial of order n of f at x = 0.

(b) Show that f (x) 6= limn Pn (x) for all x near 0. That is, the Taylor
series of f about x = 0 does not converge to f (x) for number very close to 0.

206
Solution.
(a) Using a graphing calculator we see that the derivative of f of any order
is 0 at x = 0. Hence, for all n 0 we have Pn (x) = 0 for x close to 0. Thus,
2
for x close to 0 we have f (x) = e x2 6= 0 = limn Pn (x)

Exercise 34.3
(a) Show that the above result holds for n = 0. Hint: Apply the Fundamental
Theorem of Caculus on the interval [a, x].
(b) Suppose that the result holds for up to n. That is, for any x [a, a + h]
we can estimate f (x) by Pn (x) for x near a :

f 0 (a) f 00 (a) f (n) (a)

f (x) = f (a) + (x a) + (x a)2 + + (x a)n + Rn+1 (x)
1! 2! n!
Suppose that f has continuous derivatives up to order n + 2. Use integration
by parts to show that

f (n+1) (a)
Rn+1 (x) = (x a)n+1 + Rn+2 (x).
(n + 1)!

Hence,

f 0 (a) f 00 (a) f (n+1) (a)

f (x) = f (a)+ (xa)+ (xa)2 + + (xa)n+1 +R)n + 2(x).
1! 2! (n + 1)!

Solution. Rx
(a) By the Fundamental Theorem of Calculus we have f (x) = f (a)+ a f 0 (t)dt.
Thus, the result is true for n = 0. (b) Suppose that the result holds up n.
Suppose that f (n+2) exists and continuous. Using integration by parts we
find
x
1 x (n+1) 1 x f (n+2) (t)
Z  (n+1) Z
n f (t) n+1
f (t)(x t) dt = (x t) + (x t)n+1 dt
n! a (n + 1)n! a n! a n + 1
Z x
f (n+1) (a) n+1 1
= (x a) + f (n+2) (t)(x t)n+1 dt
(n + 1)! (n + 1)! a
f (n+1) (a)
= (x a)n+1 + Rn+2 (x)
(n + 1)!

207
Exercise 34.4 (Lagranges Form of Remainder)
(a) Show that there exist x1 , x2 [a, x] such that f (n+1) (x1 ) f (n+1) (t)
f (n+1) (x2 ) for all t [a, x].
(b) Use (a) to show that
f (n+1) (x1 ) n+1 f (n+1) (x2 )
(x a) Rn+1 (x) (x a)n+1
(n + 1)! (n + 1)!
where Z x
1
Rn+1 (x) = f (n+1) (t)(x t)n dt.
n! a
(c) Show that
(n + 1)!
f (n+1) (x1 ) Rn+1 (x) f (n+1) (x2 ).
(x a)n+1
(d) Show that there is a c [a, x] such that
(n + 1)!
f (n+1) (c) = Rn+1 (x)
(x a)n+1
and therefore
f (n+1) (c)
Rn+1 (x) = (x a)n+1 .
(n + 1)!
Solution.
(a) Since f (n+1) (t) is continuous on [a, x], by Exercise 17.4 there exist x1 , x2
[a, x] such that
f (n+1) (x1 ) f (n+1) (t) f (n+1) (x2 ) for all t [a, x].
(b) This follows by multiplying the result in (a) by (x t)n and then inte-
grating from a to x.
(n+1)!
(c) This follows by multiplying the result in (b) by the ration (xa) n+1 .

(d) This follows by applying the Intermediate Value Theorem to the function
f (n+1) (t) on the interval [a, x]
Exercise 34.5 (Estimating Rn+1 (x))
Suppose that there is M > 0 such that |f (n+1) (x)| M for all x [a, a + h].
(a) Show that for all x [a, a + h] we have
M
|Rn+1 (x)| |x a|n+1 .
(n + 1)!|

208
(b) Show that
lim Rn+1 (x) = 0.
n
Hint: Exercise 1.14 and Squeeze rulw.

Solution.
(a) We have
 (n+1) 
f (c) 
n+1  M
|Rn+1 (x)| = 
 (x a)  |x a|n+1 .
(n + 1)! (n + 1)!|
(b) Since
(x a)n+1
lim =0
n (n + 1)!

and
(x a)n+1 (x a)n+1
M Rn+1 (x) M
(n + 1)! (n + 1)!
the result follows from the squeeze rule

Exercise 34.6
Let f : R R be a function such that f, f 0 , f 00 exist and are continuous.
Furthermore, f 0 and f 00 0. Show that f is a constant function.

Solution.
Let a R. Using Taylor Theorem we can write
f 00 (c)
f (x) = f (a) + f 0 (a)(x a) + (x a)2
2
for some c between a and x. But f 00 0 and f 0 so that the last equality
implies that
f 00 (c)
f (a) + f 0 (a)(x a) = f (x) (x a)2 f (x) 0
2
for all x R. But this true only when f 0 (a) = 0 and f (a) 0. Since a was
arbitrary, we conclude that f 0 (x) = 0 for all x R. Hence, f is a constant
function

Exercise 34.7
Find the Taylor polynomial of order n about 0 for f (x) = ex , and write down
the corresponding remainder term.

209
Solution.
Since the derivative of f of any order is just ex we find

x x2 xn
f (x) = 1 + + + + + Rn+1 (x)
1! 2! n!
where
ec
Rn+1 (x) = xn+1
(n + 1)!

Exercise 34.8
Find the Taylor Polynomial of order 3 for the function f (x) = cos x centered
at x = 6 .

Solution.
Simple calculation leads to

3 1  3 2 1  3
P4 (x) = x x + x
2 2 6 4 6 12 6
Exercise 34.9
Find the Lagrange form of the remainder Rn+1 (x) for the function f (x) =
1
1+x
.

Solution.
By successive differentiation we find f (n+1) (x) = (1)n+1 (n+1)!(1+x)(n+2) .
Hence,
(1)n+1 n+1
Rn+1 (x) = x
(1 + c)n+2

Exercise 34.10
Let g(x) be a function such that g(5) = 3, g 0 (5) = 1, g 00 (5) = 1 and g 000 (5) =
3.
(a) What is the Taylor polynomial of degree 3 for g(x) near 5?
(b) Use (a) to approximate g(4.9).

Solution. 00 g 000 (5)

(a) We have: c0 = g(5) = 3, c1 = g 0 (5) = 1, c2 = g 2!(5) = 12 , and c3 = 3!
=
12 . Thus, P3 (x) = 3 (x 5) + 12 (x 5)2 21 (x 5)3 .
(b) g(4.9) = 3 (4.9 5) + 21 (4.9 5)2 12 (4.9 5)3 = 3.1675.

210
Exercise 34.11
Suppose that the function f (x) is approximated near x = 0 by a sixth degree
Taylor polynomial
P6 (x) = 3x 4x3 + 5x6 .
Find the value of the following:
(a) f (0) (b) f 0 (0) (c) f 000 (0) (d) f (5) (0) (e) f (6) (0)

Solution.
If
P6 (x) = c0 + c1 x + c2 x2 + c3 x3 + c4 x4 + c5 x5 + c6 x6
then c0 = 0, c1 = 3, c2 = 0, c3 = 4, c4 = c5 = 0, and c6 = 5.

(a) f (0) = c0 = 0, (b) f 0 (0) = c1 = 3, (c) f 000 (0) = 3!c3 = 24, (d)
f (5) (0) = 5!c5 = 0, (e) f (6) (0) = 6!c6 = 3600.

Exercise 34.12
Find the third degree Taylor polynomial approximating

f (x) = arctan x,

near a = 0.

Solution.
We have

f (x) = arctan x, c0 = f (0) = 0

1
f 0 (x) = , c1 = f 0 (0) = 1
1 + x2
f 00 (0)
f 00 (x) = (1 + x2 )2 (2x), c2 = =0
2!
f 000 (0) 1
f 000 (x) = 2(1 + x2 )3 (4x2 ) 2(1 + x2 )2 , c3 = =
3! 3
Thus,
1
P3 (x) = x x3 .
3

211
Exercise 34.13
Find the fifth degree Taylor polynomial approximating

f (x) = ln (1 + x),

near a = 0.

Solution.
We have
f (x) = ln (1 + x), c0 = f (0) = 0
f 0 (x) = 1
1+x
, c1 = 0
f (0) = 1
f 00 (0)
f 00 (x) 1
= (1+x)2 , c2 = 2!
= 12
f 000 (0)
f 000 (x) = 2
(1+x)3
, c3 = 3!
= 1
3
6 f (4) (0)
f (4) (x) = (1+x) 4, c4 = 4!
= 14
24 f (5) (0) 1
f (5) (x) = (1+x)5
, c5 = 5!
= 5

Thus,
1 1 1 1
P5 (x) = x x2 + x3 x4 + x5 .
2 3 4 5

212
Solutions to Section 35
Exercise 35.1
Let f (x) = cos x.
(a) Using successive differentiation find a formula for f (n) (0).
(b) Show that
2n 2k n
1 2 1 4 n x
X
k x
P2n (x) = P2n+1 (x) = 1 x + x + (1) = (1) .
2! 4! (2n)! k=0 (2k)!

(c) Find the radius of convergence of the series

X x2n x2 x4
(1)n =1 +
n=0
(2n)! 2! 4!

(d) Show that

|x|n+1
|Rn+1 (x)| .
(n + 1)!
(e) Show that limn Rn+1 (x) = 0. Hence, conclude that

x2 x4 n x
2n
cos x = 1 + + (1) +
2! 4! (2n)!

Solution.
(a) By successive differentiation we find

f 0 (x) = sin x
f 00 (x) = cos x
f 000 (x) = sin x
f (4) (x) = cos x
..
.

We see that the derivatives go through a cycle of length 4 and then repeat
that cycle forever. It follows that

(k) 0 if k is odd
f (0) = k
(1) 2 if k is even.

213
(b) Hence,
n
1 1 x2n X x2k
P2n (x) = P2n+1 (x) = 1 x2 + x4 + (1)n = (1)k .
2! 4! (2n)! k=0 (2k)!

(c) Now, consider the series

X x2n x2 x4
(1)n =1 +
n=0
(2n)! 2! 4!

(1)n
and let an = (2n)!
. Then
 
1  an+1  1
= lim 
  = lim = 0 < 1.
R n an  n (2n + 2)(2n + 1)

This shows that the series is convergent for all values of x.

(d) Since |f (n+1) (c)| 1 we find that

|x|n+1
|Rn+1 (x)| .
(n + 1)!

(e) But
xn+1
lim =0
n (n + 1)!

so that limn Rn+1 (x) = 0. This implies that

x2 x4 x2n
cos x = 1 + + (1)n +
2! 4! (2n)!

Exercise 35.2
Let f (x) = ex .
(a) Find f (n) (0) for all n 0.
(b) Find an expression for Pn (x).
(c) Consider the series

X xn x x2
=1+ + + .
n=0
n! 1! 2!

214
Find the radius of convergence.
(d) Find an expression for Rn+1 (x) and show that

lim Rn+1 (x) = 0.

n

Hence, conclude that

x x x2 X xn
e =1+ + + = .
1! 2! n=0
n!

Solution.
(a) For all nonnegative integer k we have f (k) (x) = ex and f (k) (0) = 1.
(b) Thus, the nth Taylor polynomial is given by the expression

x x2 xn
Pn (x) = 1 + + + + .
1! 2! n!
1
(c) Let an = .
Then by the ratio test we have
n!
 
 an+1 
lim   = lim 1 = 0 < 1.
n an  n n + 1

Thus, the series xn

P
n=0 n! converges for all values of x.
(d) It remains to show that the series converges to ex . For that, we need to
use Taylor Theorem. Write f (x) = Pn (x) + En (x) where
 (n+1)
ec

f (c) 
|En (x)| =  xn+1  = |x|n+1
(n + 1)! (n + 1)!
xn
(e) Since limn n!
= 0 we find

lim Rn+1 (x) = 0.

n

Hence,

x x2 X xn
ex = 1 + + + =
1! 2! n=0
n!

Exercise 35.3
Let f (x) = ln (1 + x).
(a) Find f (n) (0) for all n 0.

215
(b) Find an expression for Pn (x).
(c) Consider the series

X xn x x2
(1)n1 =1+ + + .
n=0
n 1! 2!

Find the radius of convergence.

(d) Show that
1 |x|n+1
|Rn+1 (x)| .
|1 + c|n+1 (n + 1)!
(e) Show that
lim Rn+1 (x) = 0.
n

Hence, conclude that

X xn
ln (1 + x) = (1)n1 , 1 < x 1.
n=1
n

Solution.
(a) Taking derivatives:

f 0 (x) =(1 + x)1 f 0 (0) = 1 = 0!

f 00 (x) = (1 + x)2 f 00 (0) = 1!
f 000 (x) =2(1 + x)3 f 000 (0) = 2!
f (4) (x) = 6(1 + x)4 f (4) (0) = 3!
..
.
f (n) (x) =(1)n1 (n 1)!(1 + x)n f (n) (0) = (1)n1 (n 1)!

(b) Hence,
n
x2 x3 xn X xk
Pn (x) = x + + (1)n1 = (1)k1 .
2 3 n k=1
k

(c) Letting an = (1)n1 n1 and applying the ratio test we find that
 
1  an+1 
= lim   = 1.
R n  an 

216
 P n1 xn
Hence, the series n=1 (1) n
converges for all 1 < x < 1. By the
P (1)n1
alternating series test we know that n=1 n is convergent so the interval
of convergence of the previous series is 1 < x 1.
(d) It remains to show that the series converges to ln (1 + x). Using Taylor
theorem, we can write f (x) = Pn (x) + Rn+1 (x), where
 (n+1)
|x|n+1

f (c) n+1  1
|Rn+1 (x)| =   x  .
(n + 1)! |1 + c|n+1 (n + 1)
xn
(e) Since limn n!
= 0 we find
lim Rn+1 (x) = 0.
n

Hence,

X xn
ln (1 + x) = (1)n1 , 1<x1
n=1
n

Exercise 35.4
x
Find the Taylor series of ex
about x = 0.

Solution.
Replacing x by x in the expansion of ex we find
x2 xn
ex = 1 x + + (1)n +
2! n!
Now, multiplying both sides of this equality by x to obtain
x x 2 x3 nx
n+1
= xe = x x + + + (1) +
ex 2! n!
This series converges for all x

Exercise 35.5
1
Find the Taylor series of f (x) = 1+x2
about x = 0.

Solution.
Replacing x by x2 in Formula (1 x)1 we can write
1
2
= 1 x2 + x4 x6 + + (1)n x2n +
1+x
This series converges for 1 < x < 1.

217
Exercise 35.6
Let f (x) = sin x.
(a) Using successive differentiation find a formula for f (n) (0).
(b) Show that
n
x3 x5 (1)n 2n+1 X x2k+1
P2n (x) = P2n+1 (x) = x + + x = (1)k .
3! 5! (2n + 1)! k=0
(2k + 1)!

(c) Find the radius of convergence of the series

X (1)n 2n+1 x3 x5
x =x + + .
n=0
(2n + 1)! 3! 5!

(d) Show that

|x|n+1
|Rn+1 (x)| .
(n + 1)!
(e) Show that limn Rn+1 (x) = 0. Hence, conclude that
n
X x2k+1
sin x = (1)k .
k=0
(2k + 1)!

Solution.
(a) By successive differentiation we find
f 0 (x) = cos x
f 00 (x) = sin x
f 000 (x) = cos x
f (4) (x) = sin x
..
.
We see that the derivatives go through a cycle of length 4 and then repeat
that cycle forever. It follows that

(k) 0 if k is even
f (0) = k1
(1) 2 if k is odd.
(b) Hence, for n 1,
n
x3 x5 (1)n 2n+1 X x2k+1
P2n (x) = P2n1 (x) = x + + x = (1)k .
3! 5! (2n + 1)! k=0
(2k + 1)!

218
 (1)n
(c) Let an = (2n+1)!
. Then
 
1  an+1  1
= lim 
  = lim = 0 < 1.
R n an  n (2n + 3)(2n + 2)

(1)n 2n+1
This shows that the series
P
n=0 (2n+1)! x is convergent for all values of x.
(n+1)
(d) Since |f (c)| 1 we find that

|x|n+1
|Rn+1 (x)| .
(n + 1)!

(e) But
xn+1
lim =0
n (n + 1)!

so that
lim Rn+1 (x) = 0.
n

This implies that

X (1)n 2n+1 x3 x5
sin x = x =x + +
n=0
(2n + 1)! 3! 5!

Exercise 35.7
x
Find the MacLaurin series of 12x
.

Solution.
We have
x X
n
X
=x (2x) = 4n1 xn
1 2x n=0 n=1

which is valid for all 21 < x < 1

2

Exercise 35.8
1
Find the coefficient of (x 2)2 in the Taylor series expansion of f (x) = x
about x = 2.

Solution.
f 00 (2) 1
The coefficient is 2!
= 8

219
Exercise 35.9
2
Find the Maclaurin series for the function f (x) = x6 ex . Give your answer
in sigma notation.
Solution.
We have
6 x2 6
X (x2 )n X x2n+6
xe =x = (1)n
n=0
n! n=0
n!

Exercise 35.10
Compute each of the following sums in terms of known functions:
(1)n x4n+1
(a)
P
n=0 n!
(1)n x4n+1
(b)
P
n=1 (2n+1)!
P (1)n x6n
(c) n=0 (2n+2)!
Solution.
(1)n x4n+1 P (1)n x4n
(a)
4
= xex .
P
n=0 n!
= n=0 n!
(1)n x4n+1 (1)n (x2 )2n+1
(b) = x1
2 2
= sin xxx .
P P
n=1 (2n+1)! n=1 (2n+1)!
(1)n x6n P (1)n (x3 )2n
(c) 1
= 1cos x3
P
n=0 (2n+2)! = x6 n=1 (2n)! x6

Exercise 35.11
ex +ex
The hyperbolic cosine of x is defined to be the function cosh x = 2
. Find
the MacLaurin series of cosh x.
Solution.
Using the Taylor expansion of ex and ex one can easily see that

X x2n
cosh x =
n=0
(2n)!

Exercise 35.12
ex ex
The hyperbolic sine of x is defined to be the function sinh x = 2
. Find
the MacLaurin series of sinh x.
Solution.
Using the Taylor expansion of ex and ex one can easily see that

X x2n+1
sinh x =
n=0
(2n + 1)!

220
Exercise 35.13 (Binomial Series)
Consider the function f (x) = (1 + x)n where n R.
(a) Using successive differentiation show that f (k) (0) = k(k1) (kn+1).
(k)
Thus, f k!(0) = C(n, k) where
n!
C(n, k) = k!(nk)!
and C(n, 0) = 1.
P
(b) Find the interval of convergence of the binomial series (1+x)n = k=0 C(n, k)xk .

Solution.
(a) This is done by successive differentiation.
(b) Using the absolute ration test we find
 
 an+1  |n k|
 an  = k + 1 |x| |x| as k .
 

Thus, the interval of convergence if |x| < 1. Convergence at the endpoints

depends on the values of n and needs to be checked every time

Exercise 35.14
Find the MacLaurin series of f (x) = 1 .
x+1

Solution.
We have

12
X 1 1 (1)(3) (1)(3)(5) 3
(1 + x) = C( , k)xk = 1 x + 2 x2 x +
n=0
2 2 2 2! 23 3!

221
Solutions to Section 36
Exercise 36.1P

SupposePthat n=1 fn (x) converges uniformly on D. For each x D let

f (x) = n=1 fn (x). That is, {Sn } n=1 converges uniformly to f.
(a) Let  > 0 be given. Show that there is a positive integer N such that if
n N we have  
X n  
fk (x) f (x) <
 
 2


k=1

for all x D.
(b) Show that for n > m N we have
   
 Xn  X n m
X 
fk (x) =  fk (x) fk (x) < 
   

   
k=m+1 k=1 k=1

for all x D.

Solution.
(a) Since the series
P
n=1 fn (x) is uniformly convergent, the sequence of par-
tial sums {Sn }n=1 is uniformly convergent, say to a function f. Thus for a
given  > 0, there is a positive integer N such that for n N we have

| nk=1 fk (x) f (x)| < 2 for all x D.

P

(b) If n > m N then we have

   
 X n  X n Xm 
fk (x) =  fk (x) fk (x)
   

   
k=m+1 k=1 k=1
   
X n  X m 
 fk (x) f (x) +  fk (x) f (x)
   
   
k=1 k=1
 
< + =
2 2
Exercise 36.2 (Weierstrass)
For each integer n , let fn : D R be a continuous function that is
bounded on D with |fn (x)| Mn for all x D. Suppose that the series

222
 P
of numbers n=1 Mn is convergent. For each positive integer n define the
partial sum
n
X
Sn (x) = fk (x).
k=1

(a) Let  > 0 be given. Show that there is a positive integer N such that for
all m, n N we have  
X n Xm 
Mk Mk  < .
 

 
k=1 k=1

Hint: The sequence { nk=1 Mk }

P
n=1 is Cauchy. P
(b) Suppose that n > m N. By (a) we have | nk=m+1 Mk | < . Show that
for all x D we have
|Sn (x) Sm (x)| < 
Hence, the sequence {Sn } Pisuniformly Cauchy.
n=1
(c) Conclude that the series n=1 fn is uniformly convergent. Hint: Exercise
32.15

Solution.
(a) Since the series
P Pn
n=1 Mn is convergent, the sequence { k=1 Mk }n=1 is
convergent and hence it is Cauchy. Thus, for a fixed  we can find a positive
integer N such that for m, n N we have
 
Xn m
X 
M M k  < .
 
 k
 
k=1 k=1
Pn  Pn Pn
(b)
Pn We have |S n (x)S m (x)| = 
k=m+1 fk (x)
k=m+1 |fk (x)| k=m+1 Mk =
| k=m+1 Mk | < . Hence, the sequence {Sn } n=1 is uniformly Cauchy.
(c) This follows from Exercise 32.15

Exercise 36.3
Use Weierstrass M test to show that the series xn
P
n=0 3n
converges uniformly
on [2, 2].

Solution.  n   n
n n
Note first that
for all x [2, 2] we
have  x3n  =  x3  32 . Let fn (x) = x3n
. 2 n
and Mn = 32 The series
P
n=0 3 is a convergent geometric series. Hence,
by Weierstrass M test the given series is uniformly convergent in [2, 2]

223
Exercise
P 36.4
Let n=0 an xn be a power series with radius of convergence R. Let 0 < c < R
and D = [c, c].
(a) Define fn (x) = an xn and Mn = |an cn |.PClearly, fn is continuous in D and
Mn > 0 for all integer n 0. Show that n=0 Mn converges. Hint: Exercise
33.1(f)
(b) Let x D. Show that if x [0, c] then |gn (x)| Mn . Hint: xn is
increasing for x 0.
(c) Answer the same question if x [c, 0]. (d) Conclude that the series is
uniformly convergent on D.

Solution. P n
P n=0 an c is absolutely convergent by Exercise
(a) Since 0 < c < R the series
33.1(f). That is, the series n=0 Mn is convergent.
(b) If x [0, c] then x 0 and (xn )0 0 so that xn is increasing in [0, c].
Since 0 x c w conclude that |fn (x)| = |an ||x|n |an |cn = Mn .
(c) If x [c, 0] then x [0, c] so that |fn (x)| = |an ||x|n = |an || x|n
|an |cn = Mn .
(d) This follows from Weiestrass M test

Exercise 36.5
Show that the following series converges uniformly.

X x2
.
n=0
3n (x2 + 1)

Solution.
We have
x2 1
0 n 2
n.
3 (x + 1) 3
Since the series 1
P
n=0 3n converges, the given series converges uniformly by
the Weierstrass M test

Exercise 36.6
Let {an } Pbe
n=1 a bounded sequence with |an | M for all n N. Show that
an
ther series n=1 nx converges uniformly for all x c > 1.

224
Solution.  
We use thePWeierstrass M-test. For x c we have  nanx  M
nc
. Because c > 1,
the series M
n=1 nc converges (p-series with p > 1). Thus, by the Weierstrass
M=test, the given series converges uniformly and absolutely for x c

Exercise 36.7
Show that the series sin nx
P
n=1 n2
converges uniformly for all x R.

Solution.
We have  
 sin nx  1
0  2  2

n n
Since the series 1
P
n=1 n2 is convergent, the Weierstrass M-test asserts that
the given series converges uniformly for all x R.

Exercise 36.8
Suppose that {fn } n=1 is a sequence of functions defined on a set D
Psuch that
|fn+1 (x) fn (x)| Mn for all x
PD and n N. Assume that n=1 Mn is
convergent. Show that the series n=1 fn (x) is uniformly convergent on D.

Solution.
By the Weierstrass M-Test, we know that the series
P
n=1 (fn+1 (x) fn (x))
is uniformly convergent on D say to a function f. The nth partial sum of
this series is fn+1 (x) f1 (x). Hence, fn f + f1 uniformly on D

Exercise 36.9
Show that the series x
P
n=1 (1+x)n
converges uniformly on [1, 2].

Solution.
x
We have x 2 3x 2 + 2x 1+x
32 . Thus,
 n
x xn 2
n
n
.
(1 + x) (1 + x) 3
2 n
Since the series
P

n=1 3 converges, by the Weierstrass M-test the given
series is uniformly convergent on [1, 2]

Exercise 36.10
Prove that x
P

n=1 sin n2
converges uniformly on any bounded interval [a, b].

225
Solution.
We have   x  |x| M
sin 2  2 2
 
n n n
P M
where M = |a|+|b|. Since the series n=1 n2 is convergent, by the Weierstrass
M-test the given series is uniformly convergent

Exercise 36.11
Show that the series 1 x
P

n=1 3n
cos 3n
converges uniformly on R.

Solution.
We have
1 x 1
n
cos n
n.
3 3 3
The result now follows from Weierstrass M-test

226
Solutions to Section 37
Exercise 37.1
Let c D. Let R0 > 0 P be a number such that |c a| < R0 < R. By Exercise
36.4, the power series n
n=0 an (x a) converges uniformly on the interval
[a R0 , a + R0 ].
(a) Let  > 0 be given. Show that there is a positive integer N such that for
all n > m N we have
Pn k
Pn  Pn
k
 
k

k=0 a k (x a) k=0 a k (x a) = 
k=m+1 a k (x a) < 3 for all
x [a R0 , a + R0 ].
Hint: Exercise 36.1
(b) Show that there is a 1 > 0 such that if |x a| < 1 then
 
XN XN  
k k
ak (x a) ak (c a)  < .

 3


k=0 k=0

(c) Let = min{1 , R0 |c a|}. Show that for |x a| < we have

|f (x) f (c)| < .
P
Hence, the function f (x) = n=0 an (x a)n is continuous on D.
Solution.
(a) This follows
PN from Exercise 36.1.
k
(b) Since k=0 ak (x a) is a polynomial, it is continuous
 at c.  P
 PN k
P k N
(c) Suppose that |xa| < . Then |f (x)f (c)| =  a
k=0 k (x a) + a
k=N +1 k (x a) k=
P 
 N N
   
 k=0 ak (x a)k k=0 ak (c a)k + k=N +1 ak (x a)k + k=N +1 ak (c a)k  <
P  P P

+ 3 + 3 = . Hence, the function f (x) = n
P
3 n=0 an (x a) is continuous at
c. Since c was arbitrary, f (x) is continuous in D
Exercise 37.2
Let f (x) = n
P
n=0 an (x a) where the power R xseries converges for |x a| < R
and diverges for |x a| > R. Let F (x) = a f (t)dt. Suppose that a R <
x a. A similar result holds for a x < a + R. (a) Show that {Sn } n=1
converges uniformly
Ra to f on [x, a].
(b) Evaluate x Sn (t)dt.
an (xa)n+1
(c) Show that the power series
P
n=0 has radius of convergence R.
P an (xa)n+1 n+1
(d) Show that F (x) = n=0 n+1
. Hint: Exercise 32.11

227
Solution.
(a) See Remark ?? (2).
(b) By integration we have
Z a Z a Xn
! " n #a n
X ak (t a)k+1 X ak (x a)k+1
k
Sn (t)dt = ak (t a) dt = = .
x x k=0 k=0
k+1 k=0
k+1
x
P (xa)n+1
(c) Let R0 be the radius of convergence of n=0 an n+1 . We have
 a (xa)n+2 
1  n+1 
n+2
= lim
 
R0 n  an (xa)n+1 
 
n+1
 an+1 (x a)n+1 
 
n+1 = 1
= lim lim 
n n + 2 n an (x a)n  R
Hence, R0 = R.
(d) By Exercise R32.11 limit and
R aintegration can be interchanged. Hence,
x Ra
we have F (x) = a f (t)dt = x limn Sn (t)dt = limn x Sn (t)dt =
k+1 P an (xa)n+1
limn nk=0 ak (xa)
P
k+1
= n=0 n+1

Exercise 37.3
Let f (x) = n
P
n=0 an (x a) where the power series converges for |x a| < R
and diverges for |x a| > R. P n1
(a) Show that the power series g(x) = n=1 nan (x a) has radius of
convergence R. R
x
(b) Let G(x) = a g(t)dt. Show that G(x) = f (x) a0 for |x a| < R. Hint:
Exercise 37.2.
(c) Show that g(x) = f 0 (x) for all |x a| < R. Hint: Exercise 25.2
Solution.
(a) Let R0 be the radius of convergence of g(x). We have
 (n + 1)an+1 (x a)n 
 
1
= lim  
R0 n  nan (x a)n1 
 an+1 (x a)n+1 
 
n+1 = 1
= lim lim 

n
n n n an (x a)  R
Hence, R0 = R.
(b) Integrating term-by-term we find G(x) = n
P
n=1 an (x a) = f (x) a0 .
0 0
(c) By Exercise 25.2, we have g(x) = G (x) = f (x)

228
Exercise 37.4
Show that xn
P
n=1 n2 2n has radius of convergence 2 and show that the series
converges uniformly to a continuous function on [2, 2].
Solution.
Using the absolute ratio test we find
   2  
 an+1  n x
x
=   as n .
  
 an  n+1 2
 
2
Thus, the series converges for |x| < 2 so that the radius of convergence is 2.
Now,  n 
 x  n
  2 = 1.
 n2 2n  n2 2n n2
P 1
Since the series n=1 n2 is convergent, the given series converges uniformly
n
on [2, 2]. Since each function fn (x) = nx2 2n is continuous, by Exercise ??,
the series converges to a continuous function
Exercise 37.5
Let g(x) = sin (3x)
P
n=1 3n
. Prove that the series converges for all x R and
that g(x) is continuous everywhere.
Solution.
Since  
 sin (3x) 
0
  1
3 n  3n
and the series 1
P
n=1 3n is convergent, the given series converges uniformly
for all x R. By Exercise 37.1, g(x) is continuous everywhere
Exercise 37.6
Show that 1
P
n=1 n2 +x2
converges to a continuous function for all x R.
Solution.
1 1
Notice that x2 0 always. So n2 + x2 n2 + 1 which gives 0 < n2 +x 2 n2 .
P 1
Since n=1 n2 converges, by Weierstrass M-test, the given series converges
uniformly for all x R.
Recall a theorem saying that if a sequence of continuous function converges
uniformly, the limit function is also continuous (Exercise ??). Using this and
1
observe that fn (x) = n2 +x 2 are continuous (as denominator are non-zero), we
P 1
have n=1 n2 +x2 converges to a continuous function for all real number x

229
Exercise 37.7
Find the Taylor series about x = 0 of cos x from the series of sin x.

Solution.
d
We know that dx (sin x) = cos x, so we start the Taylor series of sin x and
differentiate this series term by term we get the series

x2 x4 x2n
cos x = 1 + + (1)n +
2! 4! (2n)!

Exercise 37.8
1
Find the Taylors series about x = 0 for arctan x from the series for 1+x2
.

Solution P
1 n 2n
Integrating term by term of the series 1+x2
= n=0 (1) x we find
Z 2n+1
dx X
n x
arctan x = = C + (1)
1 + x2 n=0
2n + 1

where 1 < x < 1. Since arctan 0 = 0 then C = 0 and therefore

x3 x5 X x2n+1
arctan x = x + = (1)n
3 5 n=0
2n + 1

Exercise 37.9
Use the first 500 terms of series of arctan x and a calculator to estimate the
numerical value of .

Solution.
Substituting x = 1 in to the series for arctan x gives
1 1 1 1
= 4 arctan 1 = 4(1 + + )
3 5 7 9
By using 500 terms of this series one finds

3.140.

Exercise 37.10 R1
Estimate the value of 0 sin (x2 )dx.

230
Solution.
The integrand has no antiderivative expressible in terms of familiar functions.
However, we know how to find its Taylor series: we know that

t3 t5
sin t = t +
3! 5!
Now if we substitute t = x2 , we have

x6 x10
sin (x2 ) = x2 +
3! 5!
In spite of the fact that we cannot antidifferentiate the function, we can
antidifferentiate the Taylor series:
R1 R 1 2 x6 x5
0
sin (x2 )dx = 0
(x 3! + 5! )dx
x3 x7 x11
= ( 3 73! + 115! )|10
1 1 1 1
= 3 73! + 115! 157! +
0.31026

231