We have been going places in the car of calculus for years, but this analysis course is about how the car actually works.
These notes may contain errors. In fact, they almost certainly do since they were just copied down by me during lectures and everyone makes mistakes when they do that. The fact that I had to
type pretty fast to keep up with the lecturer didn’t help. So obviously don’t rely on these notes. If you do spot mistakes, I’m only too happy to fix them if you email me at mdj27@cam.ac.uk with a message about them. Messages of gratitude, chocolates and job offers will also be gratefully received.
Whatever you do, don’t start using these notes instead of going to the lectures, because the lecturers don’t just write (and these notes are, or should be, a copy of what went on the blackboard) – they talk as well, and they will explain the concepts and processes much, much better than these notes will. Also beware of using these notes at the expense of copying the stuff down yourself during lectures – it really makes you concentrate and stops your mind wandering if you’re having to write the material down all the time. However, hopefully these notes should help in the following ways;

you can catch up on material from the odd lecture you’re too ill/drunk/lazy to go to; 

you can find out in advance what’s coming up next time (if you’re that sort of person) 
and the general structure of the course; 


you can compare them with your current notes if you’re worried you’ve copied something down wrong or if you write so badly you can’t read your own handwriting. 
Although if there is a difference, it might not be your notes that are wrong!
These notes were taken from the course lectured by Dr Paternain in Lent 2010. If you get a different lecturer (increasingly likely as time goes on) the stuff may be rearranged or the concepts
may be introduced in a different order, but hopefully the material should be pretty much the same. If they start to mess around with what goes in what course, you may have to start consulting the
notes from other courses. And I won’t be updating these notes (beyond fixing mistakes) – I’ll be far
too busy trying not to fail my second/third/ Good luck – Mark Jackson
th year courses.
These are the schedules for the year 2009/10, i.e. everything in these notes that was examinable in that year. The numbers in brackets after each topic give the subsection of these notes where that topic may be found, to help you look stuff up quickly.
Limits and convergence (1)
Sequences (1.1) and series (1.3) in
and
. Sums, products and quotients (1.1.2). Absolute
convergence (1.5); absolute convergence implies convergence (1.5.1). The BolzanoWeierstrass theorem (1.1.3) and applications (the General Principle of Convergence) (1.2). Comparison (1.4.1) and ratio (1.4.2) tests, alternating series test (1.4.5).
Continuity (2)
Continuity of real and complexvalued functions defined on subsets of
and
(2.1). The
intermediate value theorem (2.3). A continuous function on a closed bounded interval is bounded
and attains its bounds (2.4).
Differentiability (3)
Differentiability of functions from
to
(3.1.1). Derivative of sums and products (3.1.2). The
chain rule (3.1.3). Derivative of the inverse function (3.1.2). Rolle’s theorem (3.2.1); the mean value
theorem (3.2.2). Onedimensional version of the inverse function theorem (3.3). Taylor’s theorem from to ; Lagrange’s form of the remainder (3.5.1). Complex differentiation (3.6). Taylor’s theorem from to (statement only).
Power series (4)
Complex power series and radius of convergence (4.1). Exponential (4.3), trigonometric (4.4) and hyperbolic (4.5) functions, and relations between them. *Direct proof of the differentiability of a power series within its circle of convergence* (4.2).
Integration (5)
Definition and basic properties of the Riemann integral (5.1, 5.3). A nonintegrable function (5.1.2). Integrability of monotonic functions (5.2.2). Integrability of piecewisecontinuous functions (5.2.4). The fundamental theorem of calculus (5.4.1). Differentiation of indefinite integrals (5.4.2). Integration by parts (5.4.3). The integral form of the remainder in Taylor’s theorem (5.5.1). Improper integrals (5.6).

3 
Sequences

3 

4 
Series

5 

6 
Absolute

9 

10 

10 
11 

12 

12 

13 

13 

.............................................................................................................................. 13 

............................................................................................................ 16 

17 

18 

18 

21 

...................................................................................................................................... 21 

................................................................................................................ 21 

23 

24 

26 

28 

28 

................................................................................................................... 28 

29 

32 
5.4
Integration rules and tools ..........................................................................................................
33

.......................................................................................................... 
35 

37 
1.1 Sequences
Lemma 1.2.
as
.
Proof. The sequence Fundamental Axiom
is decreasing and bounded below . Now we claim that
.
. The
subsequence). Remark. We say nothing about uniqueness of
Proof. Let
either
and
. E.g.
. Then
1)
In case 1), set and
sequences
(i)
(ii)
(iii)
for infinitely many values of for infinitely many values of In case 2), set such that the following holds;
.
, or
for infinitely many values of
. Proceed inductively to obtain
Then
is a bounded increasing sequence and
is a bounded decreasing sequence. By the
Fundamental Axiom,
and
. By property (iii), passing to the limit,
Having selected
such that
can always do this because . Since
and
,
, select
such that
for infinitely many values of .
. We
. In other words,
This argument is called the ‘bisection method’ or ‘lion hunting’.
1.2 Cauchy sequences
Definition. A sequence . As before,
is a Cauchy sequence if given
,
Lemma 1.4. A convergent sequence is a Cauchy sequence.
Proof.
and . So if
. This means we have
,
.
. As an application of the BolzanoWeierstrass theorem, we now show that the converse is true;
Lemma 1.5. A Cauchy sequence is convergent.
Proof. First we’ll prove that if Cauchy means that , . In particular,
is a Cauchy sequence, then it is bounded. . Choose .
. Then
.
Take 
for 
. For this choice of , 
. 

Now by the BolzanoWeierstrass theorem, 
has a convergent subsequence 
. Now we 

show that in fact 
. 

. Choose 
large enough so that 

(since Cauchy, 
) and 
(since 
). Thus 
. 
Thus, for sequences of real numbers, convergence and Cauchy property are equivalent. This is called the General Principle of Convergence.
1.3 Series
1.3.1 Convergent and divergent series
We begin with some generalities.
Definition.
tends to
or
as
. We say that
converges to
. In that case we wrire
say that
diverges.
if the sequence of partial sums
. If
does not converge, we
Remark. Any question about series is really a question about the sequence of partial sums.
Lemma 1.6 (i) If (ii) Suppose
and
converge, then so does
. Then either
and
diverge (i.e. initial terms do not matter).
Proof. (i)
where both converge or both
.
If
and
, then clearly
by 1.1.
1.3.2 Convergent series and decreasing terms
Lemma 1.7. If the series
Proof.
converges, then . If the sum to infinity converges, then
. The converse is not true! Example. Consider the series
.
. Therefore
Then
, but the series diverges. This can be shown as follows;
so limit we get
. So if
converges, then
which is a contradiction.
1.3.3 Geometric series
Consider
, but since
, in the
If 
, 

If 
, the series clearly diverges. First note that 
. So, if 

, 

In conclusion, 
converges iff 
, and if 
then 
. 
1.4 Convergence tests
The first four tests in this section are for series of positive terms.
1.4.1 
Comparison test 

The most powerful test for convergence comes out straight from the fundamental axiom. 

Theorem 1.8 (Comparison test). Suppose . 
. Then if 
converges, so does 

Proof. Let 

Now 
, and the same for , so 
and 
are increasing sequences. If 

, then 
, so it follows that if 
converges, then 
and thus, since 
, 

then 
is bounded above. By the fundamental axiom, has a limit. 

Remark. Since initial terms do not matter for convergence, in the theorem it is enough to 

assume that 
. 

1.4.2 
Root and ratio tests 
We will now derive two applications of the comparison test, the root and ratio tests.
Theorem 1.9 (Root test, Cauchy). If
, suppose
as
. Then if
,
converges, and if Remark. If
,
diverges.
, the test does not give any information. In fact, later on, we’ll see examples
with 
which are convergent and others which are divergent. 

Proof. Assume that 
with 
. Take 
such that 
. Then 
means 

that given 
, 
, 
, or in other words 
. Now take 
, so that converges for
, the comparison test tells us that
. But since the geometric series converges.
Suppose now that with ; in particular, does not tend to 
. By the definition of a limit, 
, 
. 

Thus 
. So by Lemma 1.7, 
diverges. 

Example. Consider 

Since 
, by the root test, 
converges. 

Theorem 1.10 (Ratio test). Suppose 
and 
. If 
, then 
converges. If 

, then 
diverges. 

Remark. Again, the test is inconclusive if 
. 
that Proof. Suppose , we have 
with 
. Take . Now write 
. As in the root test, let 
, so 

Using the above result we get 
for 
. In other words, there is a constant 

(independent of 
) such that 
for 
. Since 
, the series converges, and by 
the comparison test, 
also converges. 

Now suppose 
with 
. From the definition of limit, 
, 

. This is saying that the sequence 
increases after . In particular, 
. But this 

clearly implies that 
, and thus 
must diverge. 
1.4.3 Some inconclusive examples Examples. (i) We know
diverges, but both tests are inconclusive.
From the root test, follows. Let
because
with
, so
by the binomial expansion. Thus
because
. [This can be proved as
Hence
.] Applying the ratio test gives
(ii) Try both tests on
so
also.
Ratio test;
Root test;
So for this series it also holds that
. However, we can show that the series converges.
Thus the series
converges, and by the comparison test,
converges.
This shows that if (iii) 
and/or 
we can’t conclude anything. 
By the ratio test, 
We conclude from the ratio test that the series converges.
1.4.4 Cauchy condensation test
Theorem 1.11 (Cauchy condensation test). Let
be a decreasing sequence of positive terms.
Then 
converges iff 
converges. 

for 
and 
. 

Proof. Note that Suppose first that converges. Then 
, which by 
is 

, i.e. 

Hence 

Multiply by 
to get 

because 
converges. Since the sequence of partial sums is bounded, 
converges. 

Suppose now, conversely, that 
converges. Then 

Thus 

because Example. Consider converges. Therefore 
is bounded in 
, so 
converges. QED. 

for 
; if , then 
. Now 
is decreasing, so let’s investigate 
. 
which is a geometric series, so converges iff
. By the condensation test,
converges iff 
. 

1.4.5 
Alternating series test 

Theorem 1.12 (Alternating series test). If 
and 
, then 

converges. 

Example. 
converges, because 
and 
is decreasing. 

Proof. 
, so 

In both expressions, all terms are 
since 
is decreasing. Thus 
is an increasing sequence 

bounded above. Now by the fundamental axiom, , . 
. But now 
. Since 

Since , given 
, 
, 
. And since 
, given 
, 

, 
. Therefore 
, 
. Hence 
converges. 

1.5 Absolute convergence 

Definition. Take convergent. 
or 
. If 
is convergent, then the sequence is called absolutely 

Note. 
is convergent but not absolutely convergent. 

1.5.1 
Absolute convergence implies convergence 
Theorem 1.13. If
is absolutely convergent, then it is convergent.
Proof. Suppose first that
. Introduce
Note that , 
, and 
and . 
Since , converges, by the comparison test must also converge. 
also converge, and since 

If , write and note that 
. So, if 

and 

converges, by the comparison test, converge absolutely, therefore converge. In other words, converge. Since and converges. and , Remark. There is an alternative ‘quick’ proof of this using Cauchy sequences. First define 
For 
, 

So 
convergent 
Cauchy 
is Cauchy 
is convergent. 
1.5.2
Conditional convergence and weird sums
Definition. If
converges but
does not, then we say that
is conditionally
convergent. The word ‘conditionally’ is used because in this case, the sum to which the series
converges is conditional on the order in which the terms are taken. E.g.
converge to different limits (
A rearrangement of Theorem 1.14. If
is
and
respectively). , a bijection, taking
.
is absolutely convergent, then every series consisting of the same term in
any order (i.e. a rearrangement) has the same sum.
Proof. We prove this for
Let
be a rearrangement of
; the extension to . Let
is an exercise.
Suppose we are given a fixed 
. Then 
such that 
contains every term in 
, so 

( ). By the fundamental axiom, 
with 
. But by symmetry, 

. 

If has any sign, consider 
and from the proof of 1.13, and , 
, . Since 

converges, both 
and 
converge, and now use the case 
to conclude that 
and
, and the result follows. QED.
2.1 Definitions and basic results
2.1.1 Two definitions of continuity
where
). Usually
and
(also applies of course to
will be some kind of interval. E.g.
Definition 1. For . 
, we say 
is continuous at 
if given any sequence 
, then 

Definition 2. For , then 
, we say 
. 
is continuous at definition). (  
if given 
, 
such that if 
and 

Proof that the two definitions are equivalent; 


D2 
D1; We know that given 
, 
such that if 
and 
, then 

, 
. Take 
with 
, then it is required to prove that 
. 

Since 
, 
. But this implies 
, i.e. 


D1 
. D2; Suppose D2 not true. Then . 
we can find 
with 

and 
. Choose 
, thus there is 
with 
and 

, but 
. So 
does not tend to 
, which 

contradicts D1. 
2.1.2
Sums, products, multiples and reciprocals
Proposition 2.1. Let , . 
, and with 
both continuous at 
. Then 
, 

for any constant is also continuous at 
are also continuous at 
. In addition, if 
, then 

Proof. This is a direct consequence of Definition 1 and Lemma 1.1 (about sequences). For 

example, to show that 
is continuous at , we take 
. Since 
and 
are 
continuous at 
, 
and 
. Lemma 1.1 implies 
; that is, 
continuous at 
. Similarly with the other claims. 

Consequence. 
is clearly continuous. By Proposition 2.1, any polynomial is continuous 

at every point of 
. Quotients of polynomials (rational functions) are continuous at every point 
where the denominator does not vanish.
Proof. Take is continuous at
,
with
. Then it is required to prove that
. Call
. Since
is continuous at
. Remark. Of course, one can prove this with the

definition.
,
. Since
2.1.4 Examples
Examples. (1)
(assuming 

and 2.2. However 

so not tend to zero. 

(2) 

For 
, 
. Also
continuous). At
is continuous by 2.1 . Take
does
, is not continuous at
, but
. Thus
is continuous (same as before). Now let
from definition of
. But
so
is continuous at
.
so
2.2 Limits of functions
. We’d like to make sense of
E.g.
but 
. Similarly if 
e.g. 
. 

Definition. 
and assume that there exists a sequence 
, 

(or 
, 
as , . (Note the point ), if given . Take may be in , 
but need not be in such that whenever .) We say that and 
, then 

. 

Remarks. 1) iff for every sequence The proof of this is exactly as the proof that Definition 1 
we have , Definition 2 from the last section. , 
. 

2) If 
, then 
iff 
is continuous at 
. There is nothing to prove here 

really, it follows straight from the definitions. 

This limit enjoys the properties which one would expect, i.e. (i) the limit is unique 

(ii) if 
as 
and 
as 
, then 
, 

and 
if We will start by looking at . 
continuous. 
2.3 Intermediate value theorem
Theorem 2.3 (Intermediate value theorem). Let
. Then
takes every value which
lies between and . 

Proof. Without loss of generality, assume 
. Take 
and let 

. Now 
, because , so 
is bounded above (by 
, in fact). Thus 

has a supremum, . 

If , then continuity of at , , 
obviously holds. If , use the following argument. (Alternatively, by as is not an upper bound for , otherwise it would contradict , so . Thus . 

Given a positive integer the definition of supremum. Then . So 
.) such that , so by continuity of . Now 

Note that , otherwise 
, but 
, . Thus, for all 
sufficiently large, 

, and 
. Again by continuity of 
, 
. Since 

, 
. In the limit, 
. Hence 
. 

Application. We can show the existence of the th root of a positive number 
, where 
is a 

positive integer. Look at 
, 
, which is continuous. We look at . By the intermediate value theorem, 
in 
. Now 
, thus
. This means that
has a positive
th root.
In fact, this loss of generality
th root is unique, since if
, so
, so
is another positive
th root with
which is a contradiction. So
, then without definitely exists.
2.4 Cofu clobi ibaatib
If 
continuous, 

Theorem 2.4. 
. 

such that such that Theorem 2.5. . In words, a continuous function on a closed bounded interval is bounded and attains its bounds. 

Note. 
on 
is not bounded. 
Proof (2.4). Suppose the statement is not true; then for any positive integer
,
. By BolzanoWeierstrass,
has a convergent subsequence
, with
. Now , which is a contradiction. 
since is continuous, so 
, but 

Proof (2.5). Let 
. Then 
by 2.4. Now 
, so by the 

definition of supremum, 
such that 
. By BolzanoWeierstrass, 

has a convergent subseequence 
, and 
. But . Similarly 

with 
. Let and the infimum. , then 
, so 
by continuity of 

Alternative proof. Let . Now consider 
. Work by contradiction and suppose 

in 
. Now 
is. By 2.4 applied to 
, 
. 

Since 
is positive, 
is continuous, since . Then 
, i.e. 
, so that 

is an upper bound for the set 
. This contradicts the definition of 
, since 
.
2.5 Inverse functions Definition.
and strictly increasing if
is said to be increasing if . Similarly for decreasing and strictly decreasing.
A function is monotone if it’s either increasing or decreasing.
Theorem 2.6. Let . Then and strictly increasing.
Proof.
be continuous and strictly increasing. Let is bijective, and the inverse
and is also continuous

is injective; If if . 
then 
. Otherwise, if 
, 
, or 


, is surjective; take . 
( 
strictly increasing.) . By the intermediate value theorem, 

So 
is bijective and has an inverse 
. 


is strictly increasing; take 
, 
, 
. If 
, then since 
is 

increasing, 
, i.e. 
. 


is continuous; given 
, let 
and 
. Then 

and 
for any 
. Take 
For this 

, if 
then 
. This was for 
, but a similar argument 
gives continuity at the end points also. QED.
3.1 Basic calculus
3.1.1 Differentiability
Let 
be a function. Mostly we’ll be looking at the case where 
is an interval on 

and 
is realvalued. 

Take 
, , 
. 

Definition. 
is differentiable at 
with derivative 
if 
is differentiable on
if it is differentiable at every point in
Remarks. (i) We could also write
.
with 
. 

(ii) Consider 

Then 
. Also 
. Thus an equivalent way of 

saying that 
is differentiable at 
with derivative is to say that there is a function such 

that 
, with 
. 
is a linear function in 
.
Other ways of writing the same thing; 
, with 
as 

. Also 
with 
. 

(iii) If 
then it is continuous at 
; suppose 

with 
is differentiable at . Then 

i.e. 
is continuous at . 

Example. Take 
, 
. If 
, 

If 
, then is differentiable, but 
. But at 
, 
is not differentiable. Does 

exist? 

is continuous at , therefore 
. 
3.1.2 Constants, sums, products and reciprocals
Proposition 3.1. (i) If 
, then is differentiable at 
with 
. 

(ii) If 
differentiable at 
, and 
. 

(iii) If 
differentiable at 
, then so is , then so is , and 
. 

(iv) If 
is differentiable at 
and 
, then is differentiable at 
and 
Remark. From (iii) and (iv) we get
Proof. (i)
(ii)
All polynomials and rational functions are differentiable. 3.1.3 Compositions – the chain rule
and 
Theorem 3.2 (Chain rule). Let is differentiable at 
. Then 
which is differentiable at 
. 
, and 
, 

Example. 
, 
, 
. Then 
, so 

. 

Proof. 
differentiable at 
means 

Similarly, differentiable at 
means 

We substitute 
to obtain 

Let 
. Then 
We need to prove that
continuous at
respectively. But now
compositions of continuous functions. Thus
. Define is continuous at
and
are because it is products, sums and . QED.
. Then
3.2 The mean value theorem
Up to now. everything has worked for functions
.
. Now we look in more detail at the case of
3.2.1 
Rolle’s theorem 

We first prove the following basic existence result; 

Theorem 3.3 (Rolle’s theorem). If , then 
continuous and differentiable on 
, and 

Proof. Let 
be the maximum value of 
in 
and 
be the minimum. (By Theorem 2.5, 

the values 
and are achieved.) Let 
. If 
, then is constant 

. Otherwise, 
or . 

Suppose 
(the proof for 
is similar). We know by Theorem 2.5 that 

. We’ll prove that 
. 

Suppose first 
. Since 
is differentiable at 
, we can write for all 

with sufficiently small. If in addition we take . Since 
, then 
, 
. Absurd 

because is the maximum of 
. 

If 
, the same argument shows that there are points at left of for which 
is strictly 

bigger than 
, which is again absurd. Thus 
. 

3.2.2 
The mean value theorem 

Theorem 3.4 (Mean value theorem). If 
continuous and differentiable on , 

then 

Remark. We can rewrite this as follows; letting where 
, 

Proof. We consider the auxiliary function 
. Let’s choose 
. 

Then 
. So 

So, for this choice of 
, 
satisfies the hypothesis of Rolle’s theorem. Hence 

. But 
, so 

3.2.3 
Corollaries 

We have the following important corollary; 

Corollary 3.5. 
continuous and differentiable. 

(i) If 
, then 
is strictly increasing. 

(ii) If 
, then 
is increasing. 

(iii) If 
, then 
is constant. 

Proof. (i) 
with 
. The mean value theorem 
for 

If 
. 

(ii) If (iii) Consider then , same argument gives for , and apply mean value theorem to get on the interval , for . 
3.3 Inverse rule (Inverse function theorem)
3.3.1 
Statement and proof 

Theorem 3.6. 
continuous and differentiable in 
with 

Let 
and 
. The function 
is a bijection and 
is 

continuous on 
and differentiable on 
with 

Proof. Since 
, then by Corollary 3.5 we know that 
is strictly increasing. 

By Theorem 2.6, 
exists and is continuous. Let 
. We need to 

prove that 
is differentiable with 

If 
(small) then there is a unique we get 
such that 
. Note; 
. Writing 

Note; if 
, then 
( 
continuous at 
). 

3.3.2 
Differentiating rational powers of 

Example. 
, 
a positive integer, is differentiable in 
( 
). Then 
and 

if The inverse rule 
. 
and 

Example. 
, any integer and 
a positive integer. We find 
using the chain 

rule, as 
; 

In other words, if 
where 
is any rational number, then 
. Later on 

we’ll define 
for 
real and discover that 
. 

3.3.3 
Some lead up to Taylor’s theorem 

with 
continuous on 
and differentiable on 
. The Mean Value 

Theorem 
for 
and 
for 

. If 
, 
We’ll see that we can choose
.
(Taylor’s theorem).
3.4
Cauchy’s mean value theorem
Theorem 3.7 (Cauchy’s mean value theorem) Suppose
differentiable on
. Then
Proof. Consider the function
such that
with
are continuous and
Then
is continuous on
and differentiable on
functions with the same property.)
. (It is just a product and sum of
because in either case, 2 columns are equal. By Rolle’s theorem
. Now expand the determinant and differentiate to see that
we want. Example. Find
gives exactly what
Apply CMVT on 
. Then 
Let 
, then 
(L’Hôpital’s rule) 
3.5 Taylor’s theorem
3.5.1 Taylor’s theorem with Lagrange’s form of the remainder Theorem 3.8 (Taylor’s theorem with Lagrange’s remainder). Suppose
to order
are continuous on
, and
exists for
and its derivatives up . Then
where
Notes. 1)
2) For
. is Lagrange’s form for the remainder. this is the mean value theorem;
Proof. Consider for
is chosen so that
We’ll apply Rolle’s theorem
. Then
. Then times. We apply it to
first, to get an
such that
Apply Rolle to
in
to get
Note that in fact from the definition of
such that we see that
.
We keep applying Rolle to get
at which
for
.
Remarks. 1)
is just to simplify matters
2) The same result holds in an interval 
. 

3) 
is the ‘Maclaurin expansion’. 

3.5.4 Application to the binomial series Application. Binomial series; Claim; 
. Then 
. 
where
is a constant that depends on
and
but not
. Therefore
as
.
3.6 Some comments on differentiability of functions
Standard properties work for both functions are differentiable.
Example.
,
and
; chain rule, the sums and products of differentiable
So
does not exist!
is differentiable with ‘real glasses’
, Complex differentiable functions are called holomorphic.
,
We will look at series of the form
4.1 Radius of convergence
4.1.1 Convergence of power series
Lemma 4.1. If 
converges, and 
, then converges absolutely. 

Proof. Since 
converges, 
as 
. In particular, there is a constant 
such 

that 
. 

since 
so 
. Thus the geometric series 
converges. By comparison, 

converges, i.e. 
converges absolutely. 
Theorem 4.2. A power series either
(i) converges absolutely for all (ii) converges absolutely for all
, or inside a circle
and diverges for all
outside it, or
(iii) converges for Definition. The circle
only.
is called the circle of convergence, and
the radius
of convergence. In case (i) we agree that 
, and in case (iii) we agree that 
. 

Proof. Let then by Lemma 4.1, 
. 
and 
converges . Now 
. If 
, 
If
is unbounded, then
and we have case (i).
if
If
is bounded, there exists a finite supremum for
, then
converges absolutely.
Choose
such that
converges absolutely.
. Then
which we call
. If
, we’ll prove that
converges, and by Lemma 4.1,
, then 
diverges. Take 
such that 
. If 

Finally we show that if converges, again by Lemma 4.1, 
converges. But this contradicts the definition of 

as supremum of 
. Thus 
diverges. 

4.1.2 Computing the radius of convergence 

The next lemma is useful in computing 
; 

Lemma 4.3. If 
as 
, then 
. 

Proof. By the ratio test we have absolute convergence if 
, i.e. 

. And if 
, then 
and 
does not tend to zero. Therefore 
. 

Remark. By the root test, if 
then 
. 

Examples. 

(geometric series). We know which does not tend to . 
, but note that if 
, the series diverges, since 

as 
? If 
, 
diverges (Example Sheet 1). 

Abel’s test; 
, , but what if , 
is bounded 

bounded in 
. Therefore 
converges for 
and 
. 

has 
but converges for every 
. 
, but on
we have divergence.
.
, Conclusion In general nothing can be said at
4.2 Differentiability of power series
Theorem 4.4. Suppose
has radius of convergence
, so that
for
with
. Then
is differentiable and
Remark. Iterate this theorem, to get that
can be differentiated infinitely many times as if it
were a polynomial. Proof (nonexaminable). We’ll use two auxiliary lemmas.
Lemma 4.5. If .
Lemma 4.6. (i)
has radius of convergence
, then so do
and
(ii) 

Proof of 4.4. By Lemma 4.5, since 
. , it defines a function 

for 
with 
has radius of convergence . Then we would like to show that 

This implies that 
is differentiable with 
. 

By Lemma 4.6, Take 
. If 
, we get 
. 

By Lemma 4.5, we know that 

converges to some number . Therefore 

Proof of 4.5. 
has radius of convergence 
. Then 
has radius of convergence 

and it is required to prove that 
. 

Take 
with 
. Choose 
. Since 
, 
as 
. In 

particular, 
. 
But
converges. Ratio test;
By comparison,
converges absolutely. Therefore
But in fact we have equality because
.
So by comparison if converges absolutely, so does 
. Therefore 
. 

What we proved also implies that Proof of 4.6. (i) also has radius of convergence of 
. 
(ii)
by the binomial theorem. Therefore
Example.
be
We saw last time that it has radius of convergence
. Then
with is differentiable and
The theorem we just proved tells us right away that
defined to
4.3 The standard functions: exponentials
General remark. Let Proof. Let
,
By the chain rule, (Corollary 3.5)
be differentiable. If . Write
. But we also see constant. Therefore
4.3.1 The exponential function
Claim.
Proof.
,
for
.
, then
is constant.
,
since
.
By the general remark above,
is constant.
From the definition of
,
, so
Now at
,
so the claim is proved.
Now we restrict
Theorem 4.7.
and prove:
(iii) By groups,
is a homomorphism.
4.3.3
functions
Definition. Let
, and
Theorem 4.9. Suppose
be any positive number. Then
and
. Then
. E.g.
.
(i)
(ii) 

(iii) 

Proof. (i) 

(ii) 

(iii) 

Let 
be a positive integer. Then 
Let 
be a positive integer, then what is 
We can now set . 
for 
. This definition agrees with the one given before for 

We can now prove that 
and 
. Define the real number 
as 

, but 
, so 
. 

Then can be rewritten as 
. Then 

for Let , by the chain rule. 

. Then 
by the chain rule. 

4.4 Trigonometric functions 

We define 

Both have infinite radius of convergence (check). 
, 
. By Theorem 4.4 they’re 

differentiable, and 
, 
. 

4.4.1 Relation to exponential function 

Now 
and 
, so 
Now from definitions, Therefore 
and 
. So 
. 

It is also obvious that 
and 
. 

Now we get, using 
, that 

for 
. 

Also, 
for 
. If 
, 
from which it follows that 

and 
. Warning: 
(or 
) are not bounded for 
, since for 

, with 
, 
4.4.2 Periodicity of the trigonometric functions
Proposition 4.10. There is a smallest positive number . 
(where 
) such that 

Proof. If 
, 
. Then 

for 
and 
(since 
and 
). Therefore 
for 

. But 
for 
is a strictly decreasing function in 
. 

We’ll prove that 
, . Then by the intermediate value theorem, 
with 

. 

where each term is 

Therefore 
. Now 
where all bracketed terms are 
for 
. For 
, 

Therefore 
. 

Corollary 4.11. 

Proof. 

But we know 
. 
Definition. We define Theorem 4.12. (i)
(ii)
(iii)
. Finally,
Proof. Immediate from addition formulae and
Note. This implies periodicity of
!
4.5 Hyperbolic functions Definition.
From this we get and
Two final remarks;
, , and identities such as
1) The other trigonometric functions ( ).
. We can check (exercise) that .
, etc) are defined in the usual way (e.g.
2) “Exponentials beat powers”, i.e.
as
integer
such that
and observe (from the definition of
for
. To see this, take a positive as a power series) that
5.1 Riemann integration
bounded, i.e.
5.1.1 Dissections
.
Definition. A dissection (or partition) of the interval
is a finite subset
of
We’ll write it as
which contains
and
.
We define the upper and lower sums of
to
,
and
, as
with respect
Clearly Lemma 5.1. If and 
are dissections with 
, then 

Proof. Suppose has one extra point, let’s say 
for some 
. 

Note that 

Then 

If now 
has more than one extra point, just do this argument for each extra point. ; we already noted this. 
The proof for Lemma 5.2. If
and
is similar to the one for the upper sums. QED. are any two dissections, then
This is a key lemma for integration.
Proof.
. QED.
, so by Lemma 5.1,
5.1.2 Definitions of integrals and integrability
Definition. The upper integral is defined as
The lower integral is
and
Remark. These numbers are welldefined because .
is bounded, so
Definition. We say that a bounded function case we write
is (Riemann) integrable if
Remark.
(or just
).
, thanks to Lemma 5.2. and
Example. (Dirichlet) A function
given by
. In this
For any
because every interval
, for any partition. Therefore
and
contains a rational number. However,
so
is not integrable.
5.2 Integrability of monotonic and continuous functions
5.2.1 Riemann’s theorem
We now prove the following useful criterion for integrability;
Theorem 5.3 (Riemann’s theorem). A bounded function .
Proof. Let’s assume first that
is integrable.
is integrable iff given
Given 
, by definition of . Consider 
, 
. By definition of . Then by Lemma 5.2, 
, 

because 
on account of being integrable. 

To prove the converse, assume that given 
, 
. Then 

by definitions of 
and 
. Since this is true for all 
. 
Therefore
is integrable. QED.
5.2.2 Integrability of monotonic functions 

We now prove that monotonic functions are integrable and continuous functions are integrable. 

Remark. Monotonic and continuous functions on 
are bounded. 

Theorem 5.4. If Proof. Assume . Then monotonic, then is increasing (same proof if But is increasing, so 
is integrable. is decreasing). Take 
to be any partition of 

Thus 

Now consider, for a positive integer, 

For this 

When , take large enough so that 

Therefore by Theorem 5.3, is integrable. QED. 

5.2.3 Uniform continuity 
Lemma 5.5 (Uniform continuity). If
if
then
is continuous, then given .
and
Remark. Continuity at a point, let’s say
, means that given
, which works for every
. The lemma is saying that we can choose a
!
Proof. Suppose claim is not true. Then there exists
such that for every
, we can find
with 
but 
. 

Take 
, to get 
with 
but 
. By Bolzano 

Weierstrass, 
. Then 
. 

Let 
, then 
, so 

On the other hand, 
. 
. continuous, therefore 
, 

. Then, passing to the limit, 
, but 
. Absurd. QED. 
5.2.4 Integrability of continuous functions
Theorem 5.6. If
continuous, then
Proof. Consider the partition
with points
is integrable.
for 
a positive integer. Then 

Let 
be given, and consider the . Then for any 
given by Lemma 5.5. Choose , Lemma 5.5 tells us that 
large enough such that . 
Therefore 
By Riemann’s criterion,
Example.
is integrable.
We claim that 
is integrable and 

Take any partition, then 
since every interval 
contains irrational 

numbers where 
is zero. Clearly 
. 
To prove the claim, it suffices to show that given
theorem and the fact that
).
(by Riemann’s
Let 
with 

. Then intervals 
be given, and take an integer and , so . So . Choose a partition with length less than . 
. Then let is a finite set with cardinality such that the points . belong to the 

First sup is 
and second sup is 
(outside 
). So 
. QED. 
5.3 Elementary properties of the integral
bounded and integrable on 
then 

(1) If 
on , then 
. 

(2) 
is integrable over 
and 
. 

(3) For any constant 
. 

(4) 
, is integrable and is integrable and . 

(5) The product 

(6) If 
is integrable. except at finitely many points in 
, then 
is integrable and 
. 

, is integrable over 
and 
and 
. 

(7) If Proof (1). If , then integral is upper sum, so; 

for any 
. Now take infimum over all 
, so 

Proof (2). Observe 

Now take any two partitions 

by Lemma 5.2. Now keep 
fixed and take infimum on all s. Then 

Now take inf over all A similar argument for the lower sums gives 
since 
integrable. 

Putting everything together we get, since 
, that 
. 

Proof (3). Exercise. Proof (4). Consider 
. Then we claim that 
is integrable. 

(this needs checking). For any partition 
, 

By Riemann, since is integrable, 
. By Riemann again and , 
is 
integrable. Now note that
since
is integrable.
. So
; since
and
are integrable by (2) and (3),
follows from (1) and the fact that 
. QED. 

Proof (5). First we’ll prove that if 
is integrable. Suppose first that 
. 

Since 
is integrable, given 
, 
is integrable, then a partition of such that 
. Let 

since 
. 

bounded, thus 
where 

is integrable, by Riemann’s theorem. 

since 
If is now arbitrary (i.e. just integrable) then is integrable. is integrable by property (4). But 

is integrable, just write 
. 

Now 
To prove that integrable 
and 
integrable 
and 
integrable, and 

we’re done. Proof (6). Let 
. Then 
for every 
except perhaps a finite set (just 

from the definition of the integral) 
must be integrable with 
. But 
is 
integrable with 
. 
Proof (7). Exercise. Convention. If 
, define 
Agree that 
. 
5.4 Integration rules and tools
How do we compute
?
5.4.1 The fundamental theorem of calculus
Let
be a bounded Riemann integrable function. For
Theorem 5.7.
Proof. For
is continuous.
define
by properties of integral, assuming is bounded so
.
. Thus
In fact for any
,
. Let
, therefore
Theorem 5.8 (Fundamental Theorem of Calculus). If in addition
is continuous, then
differentiable and 
. 

Proof. For 
, 
, we’d like to make 

small as 
. We write it as 

since 

Now 

Thus 
is
So we have proved that
Now if
, then by continuity of
,
5.4.2 Integration is the “inverse” of differentiation
Corollary 5.9. If
is continuous on
, then
Proof. From 5.8 we know that is constant. Therefore
. Therefore
This gives a way of computing
if we know a “primitive” (antiderivative) for
, that is, a
function
such that
. Primitives of continuous functions always exist (5.8) and any two
primitives differ by a constant.
such that
constant.
5.4.3 Integration by parts
Corollary 5.10. Suppose
and
exist and are continuous on
. Then
Proof. Product rule gives
. Therefore
and the rest follows immediately.
5.4.4 
Integration by substitution 

Corollary 5.11 (Integration by substitution). Let 
with , 
, 

and assume that exists and is continuous on 
. Let 
be continuous. Then 

Set , 
to get a more familiar version. 

Proof. Set 

Let 
with 
taking values in 
. Then by the chain rule, 

5.5 Taylor’s theorem revisited 

5.5.1 
Taylor’s theorem with integral form of the remainder 
We now revisit Taylor’s theorem and find an integral form for the remainder.
Theorem 5.12 (Taylor’s theorem with remainder an integral). Let , then
where
be continuous for
Remark. Note that here we assume continuity of
and not just mere existence (as in our
previous versions), so this theorem is a little weaker, but just fine for most practical purposes.
Proof. Let’s make the substitution
By integration by parts this becomes
, so that
. Then
That is,
Integrate by parts
times to get
where the last term is
. QED.
5.5.2 Link to Cauchy’s form
This integral form gives back Cauchy’s and Lagrange’s forms.
Previous remarks.
continuous, then
The mean value theorem applied to . Thus
gives
(mean value theorem for integrals). Let’s apply this to
which is Cauchy’s form of the remainder.
5.5.3 Link to Lagrange’s form
To get Lagrange’s form we use the following proposition. Suppose
, then
for some
. Then
which is Lagrange’s form of the remainder.
Now we prove the proposition we just used, namely that if
, then
such that
continuous and
If you have Proof. Let
then you get the mean value theorem.
By the Fundamental Theorem of Calculus,
and
.
Cauchy’s mean value theorem (3.7) applied to and 
gives 

Since , we simplify to get the desired result. 

5.6 Infinite integrals (Improper integrals) 
