MATH 104, HOMEWORK #12

Due Thursday, April 28

This is the last graded HW of the semester. Remember, consult the Homework Guidelines for
general instructions. Results from class, our textbook, and graded homework are fair game to use
unless otherwise specified. You may also use ungraded homework results from previous problem
sets.

GRADED HOMEWORK:

1. Ross, Exercise 33.6. Then write an actual proof of Theorem 33.5 from p. 284. (You can use
the book’s outline, but fill in the details. You may also find that things are clearer if you
reorganize the argument a bit.)

Solution. Suppose f is integrable on [a, b].

For Exercise 33.5, we need to show that M (|f |, S) − m(|f |, S) ≤ M (f, S) − m(f, S) for all
subsets S ⊆ [a, b].

An elementary but tedious proof by cases goes as follows. Note that either
M (|f |, S) = M (f, S) or M (|f |, S) = −m(f, S). Similarly, either m(|f |, S) = m(f, S) or
m(|f |, S) = −M (f, S) or m(|f |, S) = 0. Checking each of the six pairings, we find that some
pairings are incompatible, but for those which can simultaneously be true, our inequality
holds.

A slicker proof uses the book’s hint. Since f (x0 ) ≤ M (f, S) for all x0 ∈ S and f (y0 ) ≥ m(f, S)
for all y0 ∈ S, we obtain the inequality |f (x0 ) − f (y0 )| ≤ M (f, S) − m(f, S). Applying the
Triangle Inequality (subtraction version), we then have

|f (x0 )| − |f (y0 )| ≤ |f (x0 ) − f (y0 )| ≤ M (f, S) − m(f, S)

for all x0 , y0 ∈ S. In other words, M (f, S) − m(f, S) is an upper bound for the set
{|f (x0 )| − |f (y0 )| : x0 , y0 ∈ S}. I claim that actually M (|f |, S) − m(|f |, S) is the least upper
bound for this set. (It is very simple to verify that is is an upper bound, and you could
give a quick contradiction proof that it is the least upper bound.) From this we immediately
conclude M (|f |, S) − m(|f |, S) ≤ M (f, S) − m(f, S).

From here you should be able to quickly fill in details which will make the proof of Theorem
33.5 more readable.

2. Ross, Exercise 33.7 and 33.8. (Both part of 33.8 should be quick corollaries once you have
done the earlier bits.)

Solution. Let f be a bounded function on a, b], so there exists B > 0 such that |f (x)| ≤ B
for all x ∈ [a, b].
Math 104 Name:

7(a) Show that U (f 2 , P ) − L(f 2 , P ) ≤ 2B[U (f, P ) − L(f, P )] for all partitions P of [a, b].

Let S be any subset of [a, b] and let x, y ∈ S. We then have

f 2 (x) − f 2 (y) = (f (x))2 − (f (y))2

= [f (x) + f (y)] · [f (x) − f (y)]
≤ 2B · |f (x) − f (y)|
≤ 2B · [M (f, S) − m(f, S)].

Thus we see that 2B · [M (f, S) − m(f, S)] is an upper bound for the set
{f 2 (x) − f 2 (y) : x, y ∈ S}, and actually for the set {|f 2 (x) − f 2 (y)| : x, y ∈ S}. Using our
observation from Problem 2, we notice that M (f 2 , S) − m(f 2 , S) is the least upper bound of
this set. Thus

M (f 2 , S) − m(f 2 , S) ≤ 2B · [M (f, S) − m(f, S)].

If P = {a = t0 < t1 < · · · < tn = b} is a partition of [a, b], note that the inequality above
holds for all intervals S = [tk−1 , tk ]. Turning to the inequality we would like to prove, we now
have

n
X n
X
U (f 2 , P ) − L(f 2 , P ) = M (f 2 , [tk−1 , tk ]) · (tk − tk−1 ) − m(f 2 , [tk−1 , tk ]) · (tk − tk−1 )
k=1 k=1
n
X
M (f 2 , [tk−1 , tk ]) − m(f 2 , [tk−1 , tk ]) · (tk − tk−1 )
 
=
k=1
n
X
≤ 2B · [M (f, [tk−1 , tk ]) − m(f, [tk−1 , tk ])] · (tk − tk−1 )
k=1
n
X
= 2B [M (f, [tk−1 , tk ]) − m(f, [tk−1 , tk ])] · (tk − tk−1 )
k=1
= 2B[U (f, P ) − L(f, P )]

7(b) Show that if f is integrable on [a, b], then f 2 is integrable on [a, b].

Since f is integrable, it must be bounded. Choose B > 0 so that |f (x)| ≤ B for all x ∈ [a, b].

Let  > 0, which implies 2B < 0 as well. Since f is integrable, there exists a partition P of

[a, b] such that U (f, P ) − L(f, P ) < 2B . Then by part (a),

U (f 2 , P )−L(f 2 , P ) ≤ 2B·[U (f, P )−L(f, P )] < 2B· 2B = , which implies that f 2 is integrable.

8(a) Suppose f and g are integrable on [a, b]. Show that f g is integrable on [a, b].
Math 104 Name:

Note that 4f g = (f + g)2 − (f − g)2 . We know that sums and constant multiples of integrable
functions are integrable, so f + g and f − g are integrable. Then by the Exercise 33.7b above,
(f + g)2 and (f − g)2 are integrable, and taking their difference, we see that 4f g is integrable.
Since f g = 14 4f g, i.e. f g is a constant multiple of an integrable function, we see that f g is
integrable on [a, b].

8(b)Suppose f and g are integrable on [a, b]. Show that max(f, g) and min(f, g) are integrable
on [a, b].

We previously showed (in Section 17) that max(f, g) = 21 (f + g) + 21 |f − g| and min(f, g) =

1 1
2 (f + g) − 2 |f − g|. Since f and g are integrable, so are f + g and f − g (since sums and
differences of integrable functions are integrable), and then so is |f − g| (we just showed the
absolute value of an integrable function is integrable). Finally, by the Section 17 formulas,
max(f, g) and min(f, g) are both sums of integrable functions and therefore integrable on the
same interval [a, b].

Notes. The lemma used in these two problems is nice enough that I probably should have
mentioned it in class, but unless someone has a time machine, you are stuck with just these
notes. I.e. M (f, S) − m(f, S) = sup{f (x1 ) − f (x2 ) : x1 , x2 ∈ S} = sup{|f (x1 ) − f (x2 )| :
x1 , x2 ∈ S}.

3. Ross, Exercise 34.6. (Most likely you will need to use the Chain Rule. Be sure you check you
meet the conditions to do so.) Let f be continuous on R and let
Z sin x
G(x) = f (t)dt
0
for all x ∈ R. Show G is differentiable on R and find G0 (x).

Solution. First, if we set u(x) = sin x, then note that

Z u(x)
G(x) = f (u)du.
u(0)

We know that u(x) = sin x is differentiable on all of R, and its derivative u0 (x) = cos x is
continuous on all of R. Take I to be the open interval I = (−2, 2). Notice u(x) ∈ I for all I.
Thus we meet the conditions to apply our Change of Variable Theorem. We conclude that
the composition (f ◦ u)(x) = f (sin x) is continuous on all of R and

Z u(x) Z x
G(x) = f (u)du = f (sin x) cos xdx.
u(0) 0

Since (f ◦u)0 = dx
d
f (sin x) = f (sin(x)) cos(x), we are now potentially in a good position to use
the Fundamental Theorem. To verify we can do so, we just need to know that f (sin x) cos x
Math 104 Name:

is integrable. This is indeed true – since f and sin x are continuous on all of R, so is their
composition. Then since f (sin x) and cos x are both continuous, so is their product. Thus
f (sin x) cos x is continuous on all of R. Since continuous functions are integrable on all closed
intervals contained in their domains, we see that f (sin x) cos x is integrable on any interval
[a, b] ⊆ R.
Now, let’s apply the Fundamental Theorem. Let x0 ∈ R, and take [a, b] to be any closed
interval such that x0 ∈ (a, b). Set

Z x Z 0 Z x
H(x) = f (sin x) cos xdx = f (sin x) cos xdx + f (sin x) cos xdx = C + G(x)
a a 0

for some constant C.

Then the Fundamental Theorem (Part II in our book) implies that H(x) is continuous on all
of [a, b] and in particular at x0 . Furthermore, since f (sin x) cos x is continuous at x0 , we have
that H is differentiable at x0 and H 0 (x0 ) = f (sin x0 ) cos x0 . Since this holds for every x0 in
R, we see that H is continuous and differentiable on all of R, and H 0 (x) = f (sin x) cos x.

Finally, since H and G differ by a constant, we see that G is also continuous and differentiable
on all of R and G0 (x) = H 0 (x) = f (sin x) cos x.

Notes. I misspoke a bit in my parenthetical comment – you really wanted to verify you met
the conditions to undo Chain Rule, i.e. Change of Variable/ u-Substitution. Notice our proof
shows that it doesn’t really matter which constant a we have as our lower limit in FTC II.

UNGRADED HOMEWORK:

Pay special attention to starred problems; they are usually classics we will use many times, often impor-
tant theorems hidden in the exercises.

Section Exercises
33 1, 2*, 4, 9, 10, 13, 15ab
34 1, 2, 3, 5, 11, 12