Sequences & Series of

functions
We know from our study of functions that given two functions and on
some interval ,
will be continuous whenever and are,
will be differentiable whenever and are,
will be integrable whenever and are.
But what about infinite series of functions?
Do similar implications hold, and if so under which conditions?
We shall investigate these questions in this chapter.
Our first order of business in this study, is to see in which way we can
actually make sense of series of functions.
6.1. Pointwise Limits
6.1.1. Introduction. Let be a sequence of functions all defined on some
common domain . A very natural way in which we can make sense of
say and , is by requiring the existence of functions and on such that
and for all .
However as we shall see this mode of convergence is badly behaved
under limits. We shall in particular investigate each of the following
aspects:
6.1.2 Definition(Pointwise convergence of sequences)
Let be a sequence of functions defined on a common domain .
If exists (as a real number) for all , we say that the sequence
converges pointwise on . This limit defines a function on by the
equation

We write

For the infinite sum, the simplest idea is to extend the definition of
finite sum using our familiar interpretation of convergence of an infinite
series of numbers as a limit of the sequence of partial sums.
6.1.3 Definition(Pointwise convergence of series)
For each and let

If
Converges at and we write

For If the series converges for all , we say the series converges
pointwise on to the function defined by
6.1.3 Examples
Example 1(A discontinuous limit of continuous functions)
For each let .
Each of the functions is continuous on
Notice, however, that for each and yet .
This is easy to see, but it is instructive to check the details since we can
use them later to see what is going wrong in this example.
At the right-hand endpoint it is clear that, for ,.
For and , let . Then , so for
Thus
so the pointwise limit of the sequence of continuous functions is
discontinuous at .
Example 2 (The derivative of the limit is not the limit of the derivative.)
Let . Then on Now so by the previous example,
while the derivative of the limit function, , equals zero on Thus

At .
Example 3 (The integral of the limit is not the limit of the integrals.)
In this example we consider a sequence of continuous functions, each of
which has the same integral over the domain. For each let
be defined on as follows:
is linear on and on and =0 on
It is easy to verify thaton ]. Now, for each ,
But

so that the limit of the integrals is not the integral of the limit.
These examples show that the answer to each of our three questions is
negative, in general. We present some additional examples that illustrate
similar phenomena in the exercises.
6.1.4 Interchange of Limit Operations
Before turning to uniform convergence, let us first try to get an insight
into a difficulty we must overcome if we wish affirmative answers to our
questions.
If is a sequence of continuous functions converging to a function , must
be continuous? Continuity of at a point would mean that

and this would require that

Example 4 In this example we illustrate that an interchange of limit
operations may not give a correct result. Let

Viewed as a matrix,

where we are placing the entry in the row and column. For each row ,
we have so
On the other hand, for each column so
6.2. Uniform Limits
6.2.1. Introduction. In the preceding section we saw that pointwise
limits do not in general allow the interchange of limit operations.
The problem with pointwise limits is that a sequence of functions may
not converge to a limit function at the same rate at each point of its
domain.
6.2.2. Definition(Uniform convergence)
Let be a sequence of functions defined on a common domain . We say
that converges uniformly to a function on if, for every
, there exists such that for all
.
Remark We write
[unif] on or [unif] on
to indicate that the sequence converges uniformly to .
If the domain is understood from the context, we may delete explicit
reference to and write
[unif] or [unif].
Uniform convergence plays an important role in many parts of analysis.
In particular, it figures in questions involving the interchanging of limit
processes such as those we discussed in Section 6.2.
Example(Uniform vs pointwise convergence - example 1 revisited)
Let , . We observed that the sequence converges pointwise, but not
uniformly, on or on
We realized that the difficulty arises from the fact that the convergence
near is very “slow.” But for any fixed with , the convergence is uniform
on
To see this, observe that for , .
Let . Since , there exists such that if , then . Thus, if , we have

so the same that works for , also works for all

6.3.3 The Cauchy Criterion
Suppose now that we are given a sequence of functions on an interval,
and we wish to know whether it converges uniformly to some function
on . We are not told what that limit function might be.
Definition(Uniformly Cauchy sequences)
Let be a sequence of functions defined on a set .
The sequence is said to be uniformly Cauchy on if for every
there exists such that if and , then for all .
6.2.4 Testing for uniform convergence
Theorem(Cauchy Criterion For Sequences)
Let be a sequence of functions defined on a set . Then there exists a
function defined on such thatuniformly on if and only if is uniformly
Cauchy.
Theorem(Cauchy Criterion for series) Let be a sequence of functions
defined on a set . Then the series converges uniformly to some
functions on if and only if for every there is an integer so that

for all and all .

Weierstrass M-Test
It is not always easy to determine whether a sequence of functions is
uniformly convergent. In the settings of series of functions, a certain
simple test is often useful. This will certainly become one of the most
frequently used tools in your study of uniform convergence.
Theorem(M-Test) Let be a sequence of functions defined on a set and
let be a sequence of positive constants. If

and if

for each then the series

6.2.5. Illustrative examples
Example 1 Let us show that the series

converges uniformly on any interval . Our computations could be based

on the fact that the sum of this series is known to us; it is
Fix and compute

. Since
as we may choose an integer so that

Then it follows from (*) for all and all that

It follows now, by the Cauchy criterion, that the series converges

uniformly on any interval . Observe, however, that the series does not
converge uniformly on though it does converge pointwise there.
Example 2 Consider again the geometric series on the interval (as we
did in the previous Example). Then for every Since converges, by the
M-test the series converges uniformly on
6.3. Uniform Convergence &
Continuity
6.3.1. Description.
We know that the pointwise limit of a sequence of continuous
functions need not be continuous.
In this section we show that the uniform limit of a sequence of
continuous functions must be continuous and similarly that a uniformly
convergent series of continuous functions must converge to a
continuous function
6.3.2. Theorem Let be a sequence of functions defined on an interval ,
and let . If the sequence converges uniformly to some function on and
if each of the functions is continuous at , then the function is also
continuous at . In particular, if each of the functions is continuous on ,
then so too is .
6.3.3. Corollary If converges uniformly to on an interval and if each of
the functions is continuous on , then is continuous on.
6.4. Uniform Convergence & the
Integral
6.4.1. Introduction. If both and are Riemann-integrable on then so is
with

But what if

Put differently, given that when is it valid to compute term by term?

We answer this question by investigating the following topics:
6.4.2 Sequences of Continuous
Functions Around the middle of the nineteenth century, Weierstrass
showed that term by term integration is permissible when the series of
integrable functions converges uniformly.
Let us first verify this result for sequences of continuous functions.
Theorem(Integrability and sequences of continuous functions )
Suppose that that each function is continuous on and that the
convergence is uniform. Then
Corollary If an infinite series of continuous functions converges
uniformly to a function on an interval then is also continuous and

Example The geometric series converges pointwise on the interval Let .

By the M-test we see that this series converges uniformly on Each of
the terms in the sum is continuous. As a result we may apply our
theorem to integrate term by term just as we might have seen in a
calculus course. Thus
6.4.3 Sequences of Riemann Integrable Functions
In the previousTheorem we required that the functions be continuous.
Suppose we now weaken our hypotheses for these functions by
requiring only that they be integrable, but still requiring the sequence to
converge uniformly to . We note that in all respects the proof is the
same. Thus, if a uniformly convergent sequence of integrable functions
converges to an integrable function, we can integrate the sequence
term by term. Our next theorem shows that a uniform limit of
integrable functions must be integrable and so we have the following
extension of the previousTheorem
Theorem(Integrability and sequences of integrable functions)
Let be a sequence of functions Riemann integrable on an interval
Ifuniformly on then is Riemann integrable on
and
Corollary If an infinite series of integrable functions converges
uniformly to a function on an interval then is also integrable and

Example Let and for every , and so uniformly on It follows that

6.4.4. Sequences of Improper Integrals
Theorem(Dominated convergence of improper Riemann integrals)
Let be a sequence of continuous functions on the interval such
thatuniformly on any interval Suppose that there is a continuous
function g on such that

for all and all . Suppose that the integral exists. Then
6.5. Uniform Convergence &
Derivatives
6.5.1. Description. We close our analysis by showing that for functions
with continuous derivatives, uniform convergence of a sequence/series
of functions, also preserves differentiability.
6.5.1. Theorem Let be a sequence of functions each with a continuous
derivative on an interval
If the sequence of derivatives converges uniformly to a function on and
the sequence converges pointwise to a function , then f is differentiable
on and
6.5.1.Corollary Let be a sequence of functions each with a continuous
derivative on and suppose on
If the series converges uniformly on then on
Example Starting with the geometric series

we obtain from the previous Corollary that

To justify (2) we observe first that the series (1) converges pointwise on
Next we note that the series (2) converges pointwise on and uniformly
on any closed interval Thus, if and , then (2) converges uniformly on so
(2) holds at .