Chapter 1.

6, Page 37
Problem 2:
(a) Prove that x is in the Cantor set iff x has a ternary expansion that
uses only 0’s and 2’s.
(b) The Cantor-Lebesgue function is defined on the Cantor set by writ-
ing x’s ternary expansion in 0’s and 2’s, switching 2’s to 1’s, and
re-interpreting as a binary expansion. Show that this is well-defined
and continuous, F (0) = 0, and F (1) = 1.
(c) Prove that F is surjective onto [0, 1].
(d) Show how to extend F to a continuous function on [0, 1].
Solution.
(a) The nth iteration of the Cantor set removes the open segment(s) con-
sisting of all numbers with a 1 in the nth place of the ternary expansion.
Thus, the numbers remaining after n iterations will have only 0’s and
2’s in the first n places. So the numbers remaining at the end are pre-
cisely those with only 0’s and 2’s in all places. (Note: Some numbers
have a non-unique ternary representation, namely those that have a
representation that terminates. For these, we choose the infinitely re-
peating representation instead; if it consists of all 0’s and 2’s, it is in
the Cantor set. This works because we remove an open interval each
time, and numbers with terminating representations are the endpoints
of one of the intervals removed.)
(b) First, we show that this is well-defined. The only possible problem is
that some numbers have more than one ternary representation. How-
ever, such numbers can have only one representation that consists of
all 0’s and 2’s. This is because the only problems arise when one rep-
resentation terminates and another doesn’t. Now if a representation
terminates, it must end in a 2 if it contains all 0’s and 2’s. But then
the other representation ends with 12222... and therefore contains a
1.
Next we show F is continuous on the Cantor set; given  > 0, choose
k such that 21k < . Then if we let δ = 31k , any numbers within δ will
agree in their first k places, which means that the first k places of their
images will also agree, so that their images are within 21k <  of each
other.
The equalities F (0) = 0 and F (1) = 1 are obvious; for the latter,
1 = 0.2222 . . . so F (1) = 0.1111 · · · = 1.
(c) Let x ∈ [0, 1]. Choose any binary expansion of x, replace the 0’s with
2’s, and re-interpret as a ternary expansion. By part (a), this will
produce a member of the Cantor set whose image is x. (Note: Their
may be more than one preimage of x, e.g. F ( 13 ) = F ( 23 ) = 12 .)
(d) First, note that F is increasing on the Cantor set C. Now let
G(x) = sup{F (y) : y ≤ x, y ∈ C}.
Note that G(x) = F (x) for x ∈ C because F is increasing. G is
continuous at points not in C, because C̄ is open, so if z ∈ C̄, there
is a neighborhood of z on which G is constant. To show that G is
continuous on C, let x ∈ C and use the continuity of F (part b) to
1
2

choose δ > 0 such that |G(x) − G(z) <  for z ∈ C, |x − z| < δ. Choose
z1 ∈ (x − δ, x), z2 ∈ (x, x + δ) and let δ 0 < min(x − z1 , z2 − x). Then
for |y − x| < δ 0 , if y ∈ C we automatically have |F (y) − F (x)| < . If
y∈/ C but y < x, G(x) > G(y) ≥ G(z1 ) > G(x) − ; similarly, if y ∈ /C
but y > x, G(x) < G(y) ≤ G(z2 ) < G(x) + .

Problem 3: Suppose that instead of removing the middle third of the seg-
ment at each step, we remove the middle ξ, where 0 < ξ < 1.
(a) Prove that the complement of Cξ is the union of open intervals with
total length 1.
(b) Prove directly that m∗ (Cξ ) = 0.
Solution.
n
(a) At the nth
 step
n (starting at n = 0), we remove 2 segments, each of
1−ξ
length ξ 2 . The total length of these segments is
∞  n ∞
X
n 1−ξ X 1 1
2 ξ =ξ n
=ξ = 1.
n=0
2 n=0
(1 − ξ) 1 − (1 − ξ)
n
(b) If Cn is the set remaining
 nafter n iterations, then Cn is a union of 2
segments of length 1−ξ 2 . So

m(Cn ) = (1 − ξ)n .
Note that m(Cn ) → 0. Since each Cn is a covering of C by almost
disjoint cubes, the infimum of the measures of such coverings is 0.

Problem 4: Construct a closed set Ĉ so that at the kth stage of the con-
struction one removes 2k−1 centrally situated open intervals each of length
`k , with
`1 + 2`2 + · · · + 2k−1 `k < 1.
P∞
(a) If `j are chosen small enough, then k=1 2k−1 `k < 1. In this case,
show that m(Ĉ) > 0, and in fact,
∞
X
m(Ĉ) = 1 − 2k−1 `k .
k=1

(b) Show that if x ∈ Ĉ, then there exists a sequence xn such that xn ∈ / Ĉ,
yet xn → x and xn ∈ In , where In is a sub-interval in the complement
of Ĉ with |In | → 0.
(c) Prove as a consequence that Ĉ is perfect, and contains no open interval.
(d) Show also that Ĉ is uncountable.
Solution.
(a) Let Ck denote the set remaining after k iterations of this process, with
C0 being the unit segment. Then
k
X
m([0, 1] \ Ck ) = 2j `j
j=1
 3

since [0, 1] \ Ck is a union of disjoint segments with this total length.

Then
X k
m(Ck ) = 1 − 2j `j .
j=1

Now Ck & Ĉ, so by Corollary 3.3,

∞
ˆ = lim m(Ck ) = 1 −
X
m(C) 2k `k .
n→∞
k=1

(b) For k = 1, 2, . . . , let Jk be the interval of Ck which contains x. Let In

be the interval in Ĉ c which is concentric with Jn−1 . (Thus, at the nth
step of the iteration, the interval In is used to bisect the interval Jn−1 .)
Let xn be the center of In . Then xn ∈ Ĉ c . Moreover, |xn − x| ≤ |Jn−1 |
since Jn−1 contains both xn and x. Since the maximum length of the
intervals in Cn tends to 0, this implies xn → x. Finally, xn ∈ In ⊂ Ĉ c ,
and In ⊂ Jn−1 ⇒ |In | → 0.
(c) Clearly Ĉ is closed since it is the intersection of the closed sets Cn . To
prove it contains no isolated points, we use the same construction from
the previous part. Let x ∈ Ĉ. This time, let xn be an endpoint of In ,
rather than the center. (We can actually take either endpoint, but for
specificity, we’ll take the one nearer to x.) Because In is constructed
as an open interval, its endpoints lie in Cn . Moreover, successive
iterations will not delete these endpoints because the kth iteration
only deletes points from the interior of Ck−1 . So xn ∈ Ĉ. We also
have |xn − x| ≤ Jn as before, so that xn → x. Hence x is not an
isolated point. This proves that Ĉ is perfect.
(d) We will construct an injection from the set of infinite 0-1 sequences
into Ĉ. To do this, we number the sub-intervals of Ck in order from
left to right. For example, C2 contains four intervals, which we denote
I00 , I01 , I10 , and I11 . Now, given a sequence a = a1 , a2 , . . . of 0’s and
1’s, let Ina denote the interval in Cn whose subscript matches the first
n terms of a. (For instance, if a = 0, 1, 0, 0, . . . then I4a = I0100 ⊂ C4 .)
Finally, let
∞
\
x∈ Ina .
n=1
a
This intersection is nonempty because In+1 ⊂ Ina , and the intersection
of nested closed intervals is nonempty. On the other hand, it contains
only one point, since the length of the intervals tends to 0. Thus, we
have constructed a unique point in Ĉ corresponding to the sequence a.
Since there is an injection from the uncountable set of 0-1 sequences
into Ĉ, Ĉ is also uncountable.

Problem 5: Suppose E is a given set, and and On is the open set
1
On = {x : d(x, E) < }.
n
Show that
4

(a) If E is compact, then m(E) = lim m(On ).

n→∞
(b) However, the conclusion in (a) may be false for E closed and un-
bounded, or for E open and bounded.

Proof. (a) First note that for any set E,

∞
\
Ē = On
n=1

since the closure of E consists of precisely those points whose distance

to E is 0. Now if E is compact, it equals its closure, so
∞
\
E= On .
n=1

Note also that On+1 ⊂ On , so that On & E. Now since E is bounded,

it is a subset of the sphere BN (0) for some N . Then O1 ⊂ BN +1 (0),
so that m(O1 ) < ∞. Thus, by part (ii) of Corollary 3.3,
m(E) = lim m(On ).
n→∞

(b) Suppose E = Z ⊂ R. Then m(On ) = ∞ for all n, since On is a

collection of infinitely many intervals of length n2 . However, m(Z) = 0.
This shows that a closed unbounded set may not work. To construct a
bounded open counterexample, we need an open set whose boundary
has positive measure. To accomplish this, we use one of the Cantor-
like sets Ĉ from Problem 4, with the `j chosen such that m(Ĉ) > 0.
Let E = [0, 1]\ Ĉ. Then E is clearly open and bounded. The boundary
of E is precisely Ĉ, since Ĉ contains no interval and hence has empty
interior. (This shows that the boundary of E contains Ĉ; conversely,
it cannot contain any points of E because E is open, so it is exactly
equal to Ĉ.) Hence Ē = E ∪ ∂E = [0, 1]. Now
     
1 1 1 1
On = x ∈ R : d(x, E) < = x ∈ R : d(x, Ē) < = − ,1 + .
n n n n
Clearly m(On ) → 1, but m(E) = 1 − m(Ĉ) < 1.


Problem 6: Using translations and dilations, prove the following: Let B be

a ball in Rd of radius r. Then m(B) = vd rd , where vd = m(B1 ) and B1 is
the unit ball {x ∈ Rd : |x| < 1}.

Solution. Let  > 0. Choose a covering {Qj } of B1 with total volume

less than m(B1 ) + rd ; such a covering must exist because the m(B1 ) is
the infimum of the volumes of such cubical coverings. When we apply the
homothety x 7→ rx to Rd , each Qj is mapped to a cube Q0j whose side
length is r times the side length of Qj . Now {Q0j } is a cubical covering of
Br with total volume less than rd m(B1 )+. This is true for any  > 0, so we
must have m(Br ) ≤ rd m(B1 ). Conversely, if {Rj } is a cubical covering of
Br whose total volume is less than m(Br ) + , we can apply the homothety
x 7→ 1r x to get a cubical covering {Rj0 } of B1 with total volume less than
 5

1 1
rd
(m(Br )+ ). This shows that m(B1 ) ≤ rd
(m(Br ) + ). Together, these
inequalities show that m(Br ) = rd m(B1 ). 
Problem 7: If δ = (δ1 , . . . , δd ) is a d-tuple of positive numbers with δi > 0,
and E ⊂ Rd , we define δE by
δE = {(δ1 x1 , . . . , δd xd ) : (x1 , . . . , xd ) ∈ E}.
Prove that δE is measurable whenever E is measurable, and
m(δE) = δ1 . . . δd m(E).
Solution. First we note that for an open set U , δU is also open. We could
see this from the fact that x → δx is an invertible linear transformation, and
therefore a homeomorphism. More directly, if p ∈ U , let Br (p) be a neigh-
borhood of p which is contained in U ; then if we define δ̄ = min(δ1 , . . . , δd ),
we will have Bδ̄r (δp) ⊂ δU .
Next, we note that for any set S, m∗ (δS) = δ1 . . . δd m∗ (S). The proof of
this is almost exactly the same as Problem 6: the dilation x 7→ δx and
its inverse map rectangular coverings of S to rectangular coverings of δS
and vice versa; but since the exterior measure of a rectangle is just its area
(Page 12, Example 4), the infimum of the volume of rectangular coverings
is the same as the infimum over cubical coverings. Hence a rectangular
covering within  of the infimum for one set is mapped to a rectangular

covering within δ1 ...δ n
for the other.
As a more detailed version Pof the preceding argument, suppose {Qj } is a
cubical covering of SP with |Qj | < m∗ (S)+. Then {δQj } is a rectangular
covering of δS with |δQj | < δ1 . . . δd m∗ (S) + δ1 . . . P
δd . Now for each rec-
tangle δQj we can find a cubical covering {Q0jk } with k |Q0jk | < |δQj |+ 2j .
Then ∩j,k Q0jk is a cubical covering of δS with j,k |Q0jk | < δ1 . . . δd m∗ (S)+
P

(1 + δ1 . . . δd ). This implies that m∗ (δS) ≤ δ1 . . . δd m∗ (S). To get the re-

verse inequality we note that another δ-type transformation goes the other
direction, i.e. S = δ 0 (δS) where δ 0 = (1/δ1 , . . . , 1/δd ).

Now let U ⊃ E be an open set with m∗ (U \ E) < δ1 ...δ d
. Then δU ⊃ δE is
an open set. Moreover, δ(E\U ) = δE\δU , so m∗ (δU \δE) = δ1 . . . δd m∗ (U \
E) < . Hence δE is also measurable. (Alternatively, we could prove that
δE is measurable by appealing to Problem 8.) 
Problem 8: Suppose L is a linear transformation of Rd . Show that if E is a
measurable subset of Rd , then so is L(E), by proceeding as follows:
(a) Note that if E is compact, so is L(E). Hence if E is an Fσ set, so is
L(E).
(b) Because L automatically satisfies the inequality
|L(x) − L(x0 )| ≤ M |x − x0 |
for some M , we can see that L maps any√ cube of side length ` into a
cube of side length cd M `, with cd = 2 d. Now if m(E)
P = 0, there is
a collection of cubes {Qj } such that E ⊂ ∩j Qj , and j m(Qj ) < .
Thus m∗ (L(E)) ≤ c0 , and hence m(L(E)) = 0. Finally, use Corol-
lary 3.5. (Problem 4 of the next chapter shows that m(L(E)) =
| det L|m(E).)
6

Solution.
(a) Since linear transformations on finite-dimensional spaces are always
continuous, they map compact sets to compact sets. Hence, if E is
compact, so is L(E). Moreover, because Rd is σ-compact, any closed
set is the countable union of compact sets. So if
∞
\
E= Fn
n=1

where Fn is closed, then for each n we have

∞
\
Fn = Knj
j=1

where Knj is compact; then

\
E= Knj
j,n

is a countable union of compact sets. Then

\
L(E) = L(Knj )
j,n

is too, since L(Knj ) is compact. But compact sets are closed, so this
shows that L(E) is Fσ .
(b) Let x be a corner of a cube Q of side √ length `. Then every point
0
x in the cube is a distance of at most d` away from x, since this
is
√ the distance to the diagonally
√ opposite corner. Now |x − x0 | <
0 0
√d` ⇒ |L(x) − L(x )| < dM `. Now if Q is the cube of side length
2 dM√` centered at x, the points on the exterior of the cube are all at
least dM ` away from x. L(Q) ⊂ Q0 . Since a set of measure 0 has
a cubical covering with volume less than√, its image under L has a
cubical covering with volume less than 2 dM . This implies that L
maps sets of measure 0 to sets of measure 0.
Finally, let E be any measurable set. By Corollary 3.5, E = C ∩ N
where C is an Fσ set and N has measure 0. We have just shown
that L(C) is also Fσ and L(N ) also has measure 0. Hence L(E) =
L(C) ∩ L(N ) is measurable.

Problem 9: Give an example of an open set O with the following property:
the boundary of the closure of O has positive Lebesgue measure.
Solution. We will use one of the Cantor-like sets from Problem 4; let Ĉ
be such a set with m(hatC) > 0. We will construct an open set whose
closure has boundary Ĉ. Let us number the intervals involved in the Cantor
iteration as follows: If Cn is the set remaining after n iterations (with
C0 = [0, 1]), we number the 2n intervals in Cn in binary order, but with 2’s
instead of 1’s. For example, C2 = I00 ∩ I02 ∩ I20 ∩ I22 . The intervals in the
complement of Ĉ, denoted by subscripted J’s, are named according to the
intervals they bisected, by changing the last digit to a 1. For instance, in
C1 , the interval J1 is taken away to create the two intervals I0 and I2 . In
 7

the next iteration, I0 is bisected by J01 to create I00 and I02 , while I2 is
bisected by J21 to create I20 and I21 , etc.
Having named the intervals, let G = J1 ∩ J001 ∩ J021 ∩ J201 ∩ J221 ∩ . . .
be the union of the intervals in Ĉ c which are removed during odd steps of
the iteration, and G0 = [0, 1] \ (G ∩ Ĉ) be the union of the other intervals,
i.e. the ones removed during even steps of the iteration. I claim that the
closure of G is G ∩ Ĉ. Clearly this is a closed set (its complement in [0, 1]
is the open set G0 ) containing G, so we need only show that every point in
Ĉ is a limit of points in G. To do this, we first note that with the intervals
numbered as above, an interval Iabc... whose subscript is k digits long has
length less than 21k . This is so because each iteration bisects all the existing
I’s. In addition, an interval Jabc... with a k-digit subscript has length less
1
than 2k−1 because it is a subinterval of an I-interval with a (k − 1)-digit
subscript. Now let x ∈ Ĉ. Then x ∈ ∩n Cn so for each n we can find an
interval I (n) containing x which has an n-digit subscript. Let J (n) be the
J-interval with an n-digit subscript, whose first n − 1 digits match those of
I (n) . Then I (n) and J (n) are consecutive intervals in Cn . Since they both
1
have length at most 2n−1 , the distance between a point in one and a point
1
in the other is at most 2n−2 . Thus, if we let yn be a sequence such that
yn ∈ J (n) , then yn → x. Now let yn0 be the subsequence taken for odd
n, so that yn0 ⊂ G. Then we have constructed a sequence of points in G
which converge to x ∈ Ĉ.

We have shown that Ḡ = G∩ Ĉ. It only remains to show that ∂(G∩ Ĉ) =
Ĉ. Clearly ∂(G ∩ Ĉ) ⊂ Ĉ since G is open and is therefore contained in the
interior of G ∩ Ĉ. Now let x ∈ Ĉ. By the same construction as above, we
can choose a sequence yn ∈ J (n) which converges to x. If we now take the
subsequence yñ over even n, then yñ ∈ G0 and yñ → x. This proves that
x ∈ ∂(G ∩ Ĉ). Hence we have shown that G is an open set whose closure
has boundary Ĉ, which has positive measure. 

Problem 11: Let A be the subset of [0, 1] which consists of all numbers which
do not have the digit 4 appearing in their decimal expansion. Find m(A).

Proof. A has measure 0, for the same reason as the Cantor set. We can
construct A as an intersection of Cantor-like iterates. The first iterate
is the unit interval; the second has a subinterval of length 1/10 deleted,
with segments of lengths 3/10 and 6/10 remaining. (The deleted interval
corresponds to all numbers with a 4 in the first decimal place.) The next
has 9 subintervals of length 1/100 deleted, corresponding to numbers with
a non-4 in the first decimal place and a 4 in the second. Continuing, we get
closed sets Cn of length (9/10)n , with A = ∩Cn . Clearly A is measurable
since each Cn is; since m(Cn ) → 0, m(A) = 0. 

Problem 13:
(a) Show that a closed set is Gδ and an open set Fσ .
(b) Give an example of an Fσ which is not Gδ .
(c) Give an example of a Borel set which is neither Gδ nor Fσ .
8

Proof.
(a) Let U be open. As is well known, U is the union of the open rational
balls that it contains. However, it is also the union of the closed
rational balls that it contains. To prove this, let x ∈ U and r > 0 such
that Br (x) ⊂ U . Choose a rational lattice point q with |x − q| < 3r ,
and a rational d with 3r < d < 2r . Then Bd¯(q) ⊂ Br (x) ⊂ U and
x ∈ Bd¯(q), so any x ∈ U is contained in a closed rational ball within
U . Thus, U is a union of closed rational balls, of which there are only
countably many. For a closed set F , write the complement Rd \ F as
a union of rational balls Bn ; then F = ∩Bnc is a countable intersection
of the open sets Bnc , so F is Gδ .
(b) The rational numbers are Fσ since they are countable and single points
are closed. However, the Baire category theorem implies that they
are not Gδ . (Suppose they are, and let Un be open dense sets with
Q = ∩Un . Define Vn = Un \ {rn }, where rn is the nth rational in some
enumeration. Note that the Vn are also open and dense, but their
intersection is the empty set, a contradiction.)
(c) Let A = (Q ∩ (0, 1)) ∪ ((R \ Q) ∩ [2, 3]) consist of the rationals in (0, 1)
together with the irrationals in [2, 3]. Suppose A is Fσ , say A = ∪Fn
where Fn is closed. Then
(R \ Q) ∩ [2, 3] = A ∩ [2, 3] = (∪Fn ) ∩ [2, 3] = ∪(Fn ∩ [2, 3])
is also Fσ since the intersection of the two closed sets Fn and [2, 3] is
closed. But then
Q ∩ (2, 3) = ∩(Fnc ∩ (2, 3))
is Gδ because Fnc ∩ (2, 3) is the intersection of two open sets, and
therefore open, for each n. But then if rn is an enumeration of the
rationals in (2, 3), (Fnc ∩(2, 3))\{rn } is also open, and is dense in (2, 3).
Hence ∩(Fnc ∩ (2, 3)) \ {rj } is dense in (2, 3) by the Baire Category
Theorem. But this set is empty, a contradiction. Hence A cannot be
Fσ .
Similarly, suppose A is Gδ , say A = ∩Gn where Gn is open. Then
Q ∩ (0, 1) = A ∩ (0, 1) = (∩Gn ) ∩ (0, 1) = ∩(Gn ∩ (0, 1))
is also Gδ since Gn ∩ (0, 1) is the intersection of two open sets and
therefore open. But then if {qn } is an erumeration of the rationals in
(0, 1), (Gn ∩ (0, 1)) \ {qn } is open and is dense in (0, 1), so
∩ ((Gn ∩ (0, 1)) \ {qn })
must be dense in (0, 1). But this set is empty, a contradiction. Hence
A is not Gδ .


Problem 16: Borel-Cantelli Lemma: Suppose {Ek } is a countable family

of measurable subsets of Rd and that
∞
X
m(Ek ) < ∞.
k=1
 9

Let
E = {x ∈ Rd : x ∈ Ek for infinitely many k} = lim sup Ek .
• Show that E is measurable.
• Prove m(E) = 0.
Solution.
• Let
∞
[
Bn = Ek
k=n
be the set of x which are in some Ek with k ≥ n. Then x is in infinitely
many Ek iff x ∈ Bn for all n, so
∞
\ ∞ [
\ ∞
E= Bn = Ek .
n=1 n=1 k=n
This is a countable intersection of a countable union of measurable
sets, and hence isPmeasurable.
• Let  > 0. Since m(Ek ) converges, ∃N such that
∞
X
m(Ek ) < .
k=N
Then
∞ ∞
!
[ X
m(BN ) = m Ek ≤ m(Ek ) < 
k=N k=N

by subadditivity. But m(∩Bn ) ≤ m(BN ) by monotonicity, so m(E) <

 for all . Hence m(E) = 0.

Problem 17: Let {fn } be a sequence of measurable functions on [0, 1] with
|fn (x)| < ∞ for a.e. x. Show that there exists a sequence cn of positive
real numbers such that
fn (x)
→ 0 a.e. x.
cn
Solution. We are given that for each n,
∞  !
\ k
m x : |fn (x)| > =0
n
k=1

since this set is precisely the set where |f (x)| = ∞. Since these sets are
nested, this implies
 
k
lim m x : |fn (x)| > = 0.
k→∞ n
Hence, ∃cn such that
n cn o 1
m x : |fn (x)| > < n.
n 2
Define n cn o
En = x : |fn (x)| > .
n
10

1
Then m(En ) < 2n , so
 
∞ [
\ ∞
m Ej  = 0
m=1 j=m

by the Borel-Cantelli lemma. But the complement of this set consists of

precisely those points that are in finitely many En , i.e. those points for
which fnc(x)
n
is eventually less than 21n . Hence we have found a set of measure
fn (x)
0 such that cn → 0 on the complement. 
Problem 18: Prove the following assertion: Every measurable function is the
limit a.e. of a sequence of continuous functions.
Proof. Let f : R → R be measurable. (The problem didn’t specify whether
f can have ±∞ as a value, but I’m assuming not.) Let Bn = [−n, n]. Then
by Lusin’s Theorem, there exists a closed (hence compact) subset En ⊂ Bn
with m(Bn \En ) < 21n and f continuous on En . Then by Tietze’s Extension
Theorem, we can extend f to a continuous function fn on all of R, where
fn = f on En . (Explicitly, such an extension could work as follows: Define
fn : R → R by fn (x) = f (x) for x ∈ En ; for x ∈ / En , since the complement
is open, x is in some open interval (a, b) ⊂ Enc or in some unbounded open
interval (−∞, a) ⊂ Enc or (b, ∞) ⊂ Enc . Let fn (x) = f (a) + x−ab−a f (b) in the
first case and fn (x) = f (a) in the other two cases.)
I claim that fn → f almost everywhere. Suppose x is a point at which
fn 6→ f . Then x ∈ (Bnc ) ∪ (Bn \ En ) for infinitely many n since otherwise
fn (x) is eventually equal to f (x). Now a given x can be in only finitely many
Bnc , so it must be in infinitely many (Bn \ En ), i.e. x ∈ lim sup(Bn \ En ).
But lim sup(Bn \ En ) has measure 0 by the Borel-Cantelli Lemma. Hence
the set of x at which fn (x) 6→ f (x) is a subset of a set of measure 0, and
therefore has measure 0. 
Problem 20: Show that there exist closed sets A and B with m(A) = m(B) =
0, but m(A + B) > 0:
(a) In R, let A = C, B = C/2. Note that A + B ⊃ [0, 1].
(b) In R2 , observe that if A = I × {0} and B = {0} × I (where I = [0, 1]),
then A + B = I × I.
Solution.
(a) As noted, let C be the Cantor set, A = C, and B = C/2. Then A
consists of all numbers which have a ternary expansion using only 0’s
and 2’s, as shown on a previous homework set. This implies that B
consists of all numbers which have a ternary expansion using only 0’s
and 1’s. Now any number x ∈ 0, 1] can be written as a + b where
a ∈ A and b ∈ B as follows: Pick any ternary expansion 0.x1 x2 . . . for
x. Define (
2 (xn = 2)
an =
0 (else)
(
1 (xn = 1)
bn =
0 (else)
 11

Then a = 0.a1 a2 · · · ∈ A and b = 0.b1 b2 · · · ∈ B, and a + b = x.

(b) Duly noted. (I feel like I should prove something, but I’m not sure
what there is to prove here.)


Problem 21: Prove that there is a continuous function that maps a Lebesgue
measurable set to a non-measurable set.

Proof. As shown in the last homework, there is a continuous function f :

[0, 1] → [0, 1] such that f (C) = [0, 1], where C is the Cantor set. Let
V ⊂ [0, 1] be a Vitali set, which is non-measurable. Let E = f −1 (V ) ∩ C.
Then E is measurable since E is a subset of a set of measure 0. However,
f (E) = V which is not measurable. 

Problem 22: Let χ[0,1] be the characteristic function of [0, 1]. Show that
there is no everywhere continuous function f on R such that
f (x) = χ[0,1] (x) almost everywhere.

Proof. Suppose that such an f exists. Then f (1) = χ[0,1] (1) = 1. By

continuity, ∃δ > 0 such that |x − 1| < δ ⇒ |f (x) − 1| < 21 . In particular,
f (x) > 12 > 0 for x ∈ (1, 1 + δ). Thus
{x : f (x) 6= χ[0,1] (x)} ⊃ (1, 1 + δ) ⇒ m({x : f (x) 6= χ[0,1] (x)) ≥ δ > 0.


Problem 23: Suppose f (x, y) is a function on R2 that is separately contin-

uous: for each fixed variable, f is continuous in the other variable. Prove
that f is measurable on R2 .

Solution. For all x ∈ R and n ∈ N, we define Dxn to be the largest nth-order

dyadic rational less than or equal to x, i.e. Dxn = 2kn where 2kn ≤ x < k+12n .
Let fn (x, y) = f (Dxn , y). I will show that fn is measurable and fn → f
everywhere.
First we show that fn is measurable.
∞    
[ k k+1 k
{(x, y) : fn (x, y) > a} = (x, y) : n ≤ x < , f , y > a
2 2n 2n
k=−∞
∞      
[ k k+1 k
= , × y:f ,y > a .
2n 2n 2n
k=−∞

Now because f is continuous in y, {y : f ( 2kn , y) > a} is open, so the product

[ 2kn , k+1 k
2n ) × {y : f ( 2n , y) > a} is measurable. Hence {fn > a} is a countable
union of measurable sets, and thus measurable.
Next, we show that fn → f everywhere. Let  > 0. For any x, y, because f
is continuous in x, there is some 1-dimensional neighborhood (x ± δ, y) on
which |f (x0 , y) − f (x, y)| < . Then for sufficiently large n, |Dxn − x| < δ ⇒
|fn (x, y) − f (x, y)| = |f (Dxn , y) − f (x, y)| < . Hence fn (x, y) is within  of
f (x, y) for sufficiently large n. Since f is the pointwise limit of measurable
functions fn , it is measurable. 
12

Problem 27: Suppose E1 and E2 are a pair of compact sets in Rd with

E1 ⊂ E2 , and let a = m(E1 ) and b = m(E2 ). Prove that for any c with
a < c < b, there is a compact set E with E1 ⊂ E ⊂ E2 and m(E) = c.

Solution. Since E1 is measurable, there is an open set U ⊃ E1 with m(U \

E1 ) < b − c. Then E2 ∩ U c is compact (since it’s the intersection of a
compact set and a closed set) and has measure at least m(E1 ) − m(U ) >
b − (a + b − c) = c − a. If we can find a compact subset K ⊂ E2 ∩ U c with
m(K) = c − a, then K ∪ E1 will be a compact subset of E2 with measure
(c − a) + a = c (since K and E1 are disjoint). Hence we have reduced the
problem to the following: Given a compact set F ⊂ Rd with m(K) = µ,
and given ξ with 0 < ξ < µ, find a compact subset F 0 ⊂ F with m(F 0 ) = ξ.
This can be solved as follows: Let f (y) = m(F ∩ By (0)). Then f (0) = 0
whereas f (y) = µ for sufficiently large y (because F is bounded). Moreover,
f is continuous: Given y and  > 0, the continuity of m(By (0)) allows us
to find δ such that |y 0 − y| < δ ⇒ |m(By0 (0)) − m(By (0))| < . Then
|f (y 0 ) − f (y)| = m(F ∩ (By0 (0)∆By (0))) < , because this is the measure
of a subset of the symmetric difference (By0 (0)∆By (0)) which has measure
|m(By0 (0)) − m(By (0))| < . Hence f is continuous, so by the Intermediate
Value Theorem there is a value y0 such that f (y0 ) = ξ. Then the compact
set F ∩ By0 (0) has measure ξ as desired. 

Problem 28: Let E ⊂ R with m∗ (E) > 0. Let 0 < α < 1. Then there exists
an interval I ⊂ R such that m∗ (I ∩ E) ≥ αm(I).
P
Proof. For any , we can find a cubical covering {Qj } of E with |Qj | <
m∗ (E) + . Then, by expanding each cube to an open P cube of size 2j more,
we can construct an open cubical covering {Ij } with |Ij | < m∗ (E) + 2.
(We name these Ij because 1-dimensional cubes are in fact intervals.) Then
[ [ X
E⊂ Ij ⇒ E = (E ∩ Ij ) ⇒ m∗ (E) ≤ m∗ (E ∩ Ij ).
j j j

Now we apply something like the Pigeonhole Principle: Suppose m∗ (E ∩

Ij ) < αm(Ij ) for all j. Then
X X
m∗ (E) ≤ m∗ (E ∩ Ij ) < α m(I)j < α(m∗ (E) + 2).
j j

But if  is chosen small enough, this is a contradiction; explicitly, we can

choose  < (1−α)m 2
∗ (E)
. Thus there must be some interval Ij for which
m∗ (E ∩ Ij ) ≥ αm(Ij ). 

Problem 29: Suppose E is a measurable subset of R with m(E) > 0. Prove

that the difference set of E
{z ∈ R : z = x − y for some x, y ∈ E}
contains an open interval centered at the origin.

Solution. By Problem 28, there exists an interval I such that m(E ∩ I) ≥

3
4 m(I). (We can replace m∗ by m here because E is measurable.) Let
d = m(I) and 0 < |α| < d4 . Then the translated set I + α intersects I in an
 13

interval I 0 of length d−α > 34 d. Now m(E∩I) ≥ 34 d so m((E+α)∩(I+α)) ≥

3
4 d by the translation invariance of Lebesgue measure. Now
3
d ≤ m(E ∩I) = m(E ∩I 0 )+m(E ∩(I \I 0 )) ≤ m(E ∩I 0 )+m(I \I 0 ) = m(E ∩I 0 )+α
4
so m(E∩I 0 ) ≥ 3d d 0 d
4 −α > 2 . Similarly, m((E+α)∩I ) > 2 . Now if (E+α) and
0 0
E were disjoint, this would imply m(I ) ≥ m(E ∩ I ) + m((E + d4 ) ∩ I 0 ) > d.
But m(I 0 ) = d − α, so E and E + α must have nonempty intersection. Let
x ∈ E ∩(E +α). Then ∃e1 , e2 ∈ E such that e1 = x = e2 +α ⇒ e1 −e2 = α.
Hence α ∈ E − E for all α ∈ (− d4 , d4 ). 
Problem 30: If E and F are measurable, and m(E) > 0, m(F ) > 0, prove
that
E + F = {x + y : x ∈ E, y ∈ F }
contains an interval.
Solution. We follow the preceding proof almost exactly. By the lemma,
there exist intervals I1 and I2 such that m(E ∩ I1 ) ≥ 43 m(I1 ) and m(−F ∩
I2 ) ≥ 43 m(I2 ). WLOG assume m(I2 ) ≤ m(I1 ). Then ∃t0 such that I2 +t0 ⊂
I1 . Let d = m(I2 ). Then for 0 < |α| < d4 , I2 + t0 + α intersects I1 in
an interval I 0 of length at least d − α > 3d 4 . By the same argument as
in problem 29, this implies that I1 and I2 + t0 + α must have nonempty
intersection; let x be a point of this intersection. Then ∃e ∈ E, f ∈ F such
that e = x = −f + t0 + α ⇒ e + f = t0 + α. Hence E + F contains the
interval (t0 − d4 , t0 + d4 ). 
Problem 34: Given two Cantor sets C1 , C2 ⊂ [0, 1], there exists a continuous
bijection f : [0, 1] → [0, 1] such that f (C1 ) = C2 .
Proof. Any Cantor set can be put in bijective correspondence with the set
of 0-1 sequences as follows: Given x ∈ C, where C = C1 ∩ C2 ∩ . . . is a
Cantor set, define x1 = 0 if x is in the left of the two intervals in C1 (call
this left interval I0 ), and x1 = 1 if x is in the right interval I1 . Then define
x2 = 0 if x is in the left subinterval (either I00 or I10 ) in C2 , and x2 = 1 if x
is in the right subinterval. Continuing in this fashion, we obtain a bijection
from C to the set of 0-1 sequences. Note that this bijection is increasing in
the sense that if y > x for x, y ∈ C, then yn > xn at the first point n in the
sequence at which xn and yn differ.
Now we can create an increasing bijection f from C1 to C∈ by mapping from
C1 to 0-1 sequences, and from there to C2 . This function will be continuous
on C1 because if x, y ∈ C1 are close, their corresponding sequences {xn }
and {yn } will agree in their first N terms; then f (x) and f (y) will agree in
their first N terms as well, which means they’re in the same subinterval of
the N th iterate of C2 , which has length at most 21N . Hence f (x) and f (y)
can be made arbitrarily close if x and y are sufficiently close. Then since
C1 is compact, we can extend f to a continuous bijection on all of [0, 1] in a
piecewise linear fashion, because C1c is a disjoint union of open intervals on
which f can be made piecewise linear. This construction will also preserve
the bijectivity of f . Hence we have a continuous bijection f : [0, 1] → [0, 1]
with f (C1 ) = C2 . 
14

Problem 35: Give an example of a measurable functions f and a continuous

function Φ so that f ◦ Φ is non-measurable. Use the construction in the
hint to show that there exists a Lebesgue measurable set that is not a Borel
set.
Solution. We will make use of Problem 34. Let C1 and C2 be two Cantor
subsets of [0, 1] such that m(C1 ) > 0 and m(C2 ) = 0. Let N ⊂ C1 such
that N is non-measurable. (Here we use the fact that every set E ⊂ R
of positive measure has a non-measurable subset. This is easy to prove
by mimicking the Vitali construction but restricting it to E: Define the
equivalence relation on E by x y if x − y ∈ Q, and let N ⊂ E con-
tain one member from each equivalence class. Then N and its countably
many translates are all of E, so N cannot have either measure 0 or nonzero
measure.) Define f = χΦ(N ) . Then f ◦ Φ = χN is non-measurable, since
{x : χN (x) > 12 } = N .
To show that there is a Lebesgue-measurable set which is not Borel mea-
surable, we will use (without proof) the fact that every Borel set can be
represented by a finite number of union and intersection signs, followed by
some open sets. We also use the general fact that for any function φ and
any sets Xα , φ−1 (∩Xα ) = ∩φ−1 Xα and φ−1 (∪Xα ) = ∪φ−1 Xα . Together,
these imply that for a continuous function Φ, Φ−1 of a Borel set is Borel,
because
[ \ \  [ \ \
Φ−1 · · · Gn = · · · Φ−1 (Gn )

and Φ−1 (Gn ) is open for continuous Φ and open Gn . (Of course, the string
of cups and caps in this equation could just as well start with a cap.) Now
consider the set Φ(N ) from our construction above. Since Φ(N ) is a subset
of the set C2 which has measure 0, Φ(N ) is measurable by the completeness
of Lebesgue measure. However, Φ is a bijection so Φ−1 (Φ(N )) = N which
is not Borel, so Φ(N ) cannot be Borel. 

1. Chapter 1.7, Page 46

Problem 2: Any open set U can be written as the union of closed cubes, so
that U = ∪Qj with the following properties:
(i) The Qj have disjoint interiors.
(ii) d(Qj , U c ) ≈ side length of Qj . This means that there are positive
constants c and C so that c ≤ d(Qj , U c )/`(Qj ) ≤ C, where `(Qj )
denotes the side length of Qj .

Proof. Let U ⊂ Rd be an open set. Let P0 = {Qj } be the partition of U

d(Qj ,U c )
into dyadic cUbes as described in the book (pp. 7-8). I claim that `(Q j)
is bounded above for this partition (i.e. does
√ not take on arbitrarily large
valUes). This is so because if d(Qj , U c ) > d`(Qj ), then Qj is a subset of
a larger dyadic cube lying within U . (If we take the dyadic √ cube one size
larger than Qj and containing Qj , then Qj is at most d`Qj away from
any point in this larger dyadic cube.) But the constrUction of P0 is done
in such a way that every dyadic cube is maximal in U , i.e. √ is not contained
d(Qj ,U c )
in a larger dyadic cube lying within U . Hence `Qj ≤ d for all Qj .
 15

Now we define an iterative refining procedUre on the partition P0 . In√ this

c
procedure, every cube Q in the partition Pn for which d(Q,U `(Q)
)
< 4d is
replaced by its 2d dyadic sub-cubes. If R is one of these sub-cubes, then
d(R, U c ) d(Q, U c ) d(Q, U c )
≥ =2
`(R) `(Q) `(Q)
and √
d(R, U c ) d + d`(R) d(Q, U c ) √
≤ =2 + d.
`(R) `(R) `(Q)
d(Q,U c )
ThUs, if ρ(Q) = `(Q) ,then the sub-cubes have the property
√
2ρ(Q) ≤ ρ(R) ≤ 2ρ(Q) + d.
Now let
∞ \
[ ∞
P= Pk
n=1 k=n
be the partition consisting of those cubes that are eventUally in all the Pn .
√ c √
I claim that 4d ≤ d(Q,U`(Q)
)
≤ 2 d for any cube Q ∈ P. Consider what
happens as oUr refinement process iterates. If a given cube has too small
a distance-to-side ratio, its sub-cubes will have this ratio at least doubled
in the next iteration. Hence, after enough iterations its sub-cubes will all
√ √
have their distance-to-side ratio in the desired interval [ 4d , 2 d]. Once
they are in this interval, they are not sub-divided any further. One the
other hand, none can ever achieve a ratio too large to√be in this interval,
since a cube is only subdivided if its ratio
√ is less √ than √d/4, and√the next
iteration can then make it at most 2( d/4) + d = 3 d/2 < 2 d. This
shows that P has all its cubes in the desired interval. Consider also that
P must have disjoint interiors and cover all of U : The only cubes with
overlapping interiors are those from distinct steps in our iteration scheme,
so taking the intersection to get P will weed out any such overlaps. Also,
for a given x ∈ U , if we consider the sequence of cubes containing x in
our various partitions, this sequence will shrink for a finite number of steps
and then stay constant once the distance-to-side ratio reaches a desirable
number. Hence x is contained in some cube that is eventually in all the Pn ,
so x is covered by P. 
Problem 3: Find an example of a measurable subset C of [0, 1] such that
m(C) = 0, yet the difference set of C contains a non-trivial interval centered
at the origin.
Solution. Let C be the Cantor middle-thirds set. Note that the Cantor
dust described in the hint consists precisely of those point (x, y) for which
both x and y are in C. Note also that if a line of slope 1 passes through
any cube in any iteration of the Cantor dust, it must pass through one
of the sub-cubes of that cube in the next iteration. To see this, consider
WLOG the original cube. For a line y = x + a, if 31 ≤ a ≤ 1 then the line
passes through the upper left cube; if − 13 ≤ a ≤ 13 then it passes through
the lower left cube; and if −1 ≤ a ≤ − 31 it passes through the lower right
cube. Thus, if Cn is the nth iteration of the Cantor dust and L is a line
16

of the form y = x + a with −1 ≤ a ≤ 1, then L ∩ Cn is nonempty and

compact. (It is nonempty by the preceding remark, and compact because
it is the intersection of the compact sets Cn and L ∩ ([0, 1] × [0, 1]).) Thus,
the infinite intersection of these nested compact sets is nonempty, i.e. L
intersects the Cantor dust at some point (x, y). Then since x, y ∈ C, we
have x − y = a. So a is in the difference set of the Cantor set for any
a ∈ [−1, 1]. 

Problem 4: Complete the following outline to prove that a bounded function

on an interval [a, b] is Riemann integrable if and only if its set of disconti-
nuities has measure zero. This argument is given in detail in the appendix
to Book I.
Let f be a bounded function on a compact interval J, and let I(c, r) de-
note the open interval centered at c of radius r > 0. Let osc(f, c, r) =
sup |f (x) − f (y)| where the supremum is taken over all x, y ∈ J ∩ I(c, r),
and define the oscillation of f at c by osc(f, c) = limr→0 osc(f, c, r). Clearly,
f is continuous at c ∈ J if and only if osc(f, c) = 0.
Prove the following assertions:
(a) For every  > 0, the set of points c ∈ J such that osc(f, c) ≥  is
compact.
(b) If the set of discontinuities of f has measure 0, then f is Riemann
integrable.
(c) Conversely, if f is Riemann integrable on J, then its set of discontinu-
ities has measure 0.

Solution.
(a) Clearly this set is bounded, so we need only show that it’s closed.
Suppose c1 , c2 , . . . are a sequence of such points with cn → c. We wish
to show that osc(f, c) ≥  as well. Let r > 0. Since cn → c, there is
some cN with |cN − c| < 2r . Then since osc(f, cN ) ≥ , there must be
x, y ∈ J within 2r of cN such that |f (x) − f (y)| ≥ . But then x and
y are within r of c, so osc(f, c, r) ≥ . Since this is true for all r, we
must have osc(f, c) ≥ .
(b) Let  > 0. Let
A = {c ∈ J : osc(f, c) ≥ }.
Since A is a subset of the discontinuity set of f , it has measure 0.
Hence there it can be covered by an open set U with m(U ) < . Since
every open subset of R P is a countable disjoint union of intervals, we
can write U = ∪In with |In | < . Now because A is compact, there
is a finite subcover; by re-ordering the intervals we may write
N
[
A ⊂ In .
n=1

Now on the compact set J 0 = J \ ∪In , osc(f, c) <  for all c. For
each x ∈ J 0 , this means we can find rx such that osc(f, x, rx ) < .
Then J 0 is covered by the open intervals Ux = (x − rx , x + rx ). Let
δ be the Lebesgue number of this covering, so that any subinterval of
J 0 with length at most δ must be contained in one of the Ux . Now
 17

consider a partition of J with mesh size less than δ. The total length of
all subintervals which intersect ∪In is at most  + 2N δ since enlarging
each In by δ will cover all such intervals. On each of these subintervals,
sup f − inf f ≤ 2M where |f | ≤ M on J. Hence the contribution these
intervals make to the difference U (P, f )−L(P, f ) is at most 2M +4M δ.
The other subintervals are contained in J 0 and by construction of δ,
each is contained within some Ux , so sup f − inf f ≤  on each of them.
Hence the total contribution they make to U (P, f ) − L(P, f ) is at most
m(J). Thus, we have
U (P, f ) − L(P, f ) ≤ (2M + m(J) + 4δ).
By requiring δ to be less than some constant times , we have thus
shown that the difference between upper and lower sums can be made
smaller than a constant times . Hence f is Riemann integrable.
(c) Suppose f is Riemann integrable, and let  > 0. Let n ∈ N. Then
there is a partition P of J with U (P, f ) − L(P, f ) < n . Now if the
interior of any subinterval Ik of this partition intersects A1/n at some
x, then sup f − inf f ≥ n1 on Ik because osc(f, x, r) ≥ n1 for all r, and
(x − r, x + r) ⊂ Ik for sufficiently small r. So the total length of the
subintervals whose interiors intersect A1/n is at most  since otherwise
they would make a contribution of more than /n to U (P, f ) − (P, f ).
Hence we have covered A1/n by a collection of intervals of total length
less than , which implies m(A1/n ) < . Now if A is the set of points
at which f is discontinuous, then
[∞
A= A1/n .
n=1
Since An ⊂ An+1 , the continuity of measure implies that
m(A) = lim m(A1/n ) ≤ .
n→∞

Since m(A) ≤  for all , m(A) = 0.


Problem 6: The fact that the axiom of choice and the well-ordering principle
are equivalent is a consequence of the following considerations.
One begins by defining a partial ordering on a set E to be a binary relation
≤ on the set E that satisfies:
(i) x ≤ x for all x ∈ E.
(ii) If x ≤ y and y ≤ x, then x = y.
(iii) If x ≤ y and y ≤ z, then x ≤ z.
If in addition x ≤ y or y ≤ x whenever x, y ∈ E, then ≤ is a linear ordering
of E.
The axiom of choice and the well-ordering principle are then logicall equiva-
lent to the Hausdorff maximal principle: Every non-empty partially ordered
set has a (non-empty) maximal linearly ordered subset. In other words, if
E is partially ordered by ≤, then E contains a non-empty subset F which
is linearly ordered by ≤ and such that if F is contained in a set G also
linearly ordered by ≤, then F = G.
An application of the Hausdorff maximal principle to the collection of all
18

well-orderings of subsets of E implies the well-ordering principle for E.

However, the proof that the axiom of choice implies the Hausdorff maximal
principle is more complicated.

Solution. I don’t like this problem for two reasons: (1) It’s not very clearly
stated what exactly I’m supposed to do. My best guess is that I’m supposed
to use the axiom of choice to prove the Hausdorff maximal principle, but
the problem never really comes out and says that. (2) What the crap. This
is supposed to be an analysis course, not a set theory course. We haven’t
defined any of the basic concepts of set theory, yet I’m supposed to come
up with a “complicated” set theory proof. Since I’ve never had a course in
set theory, I really don’t feel like guessing my way to something that might
be a proof. So, here’s the proof from the appendix to Rudin’s Real and
Functional Analysis:
For a collection of set F and a sub-collection Φ ⊂ F, call Φ a subchain of
F if Φ is totally ordered by set inclusion.
Lemma 1. Suppose F is a nonempty collection of subsets of a set X such
that the union of every subchain of F belongs to F. Suppose g is a function
which associates to each A ∈ F a set g(A) ∈ F such that A ⊂ g(A) and
g(A) \ A consists of at most one element. Then there exists an A ∈ F for
which g(A) = A.
Proof. Let A0 ∈ F. Call a subcollection F 0 ⊂ F a tower if A0 ∈ F 0 , the
union of every subchain of F 0 is in F 0 , and g(A) ∈ F 0 for every A ∈ F 0 .
Then there exists at least one tower because the collection of all A ∈ F
such that A0 ⊂ A is a tower. Let F0 be the intersection of all towers,
which is also a tower. Let Γ be the collection of all C ∈ F0 such that every
A ∈ F0 satisfies either A ⊂ C or C ⊂ A. For each C ∈ Γ, let Φ(C) be the
collection of A ∈ F0 such that A ⊂ C or g(C) ⊂ A. We will prove that Γ is
a tower. The first two properties are obvious. Now let C ∈ Γ, and suppose
A ∈ Φ(C). If A is a proper subset of C, then C cannot be a proper subset
of g(A) because then g(A) \ A would have at least two elements. Hence
g(A) ⊂ C. If A = C, then g(A) = g(C). The third possibility for A is that
g(C) ⊂ A. But since A ⊂ g(A) this implies g(C) ⊂ g(A). Thus, we have
shown that g(A) ∈ Φ(C) for any A ∈ Φ(C), so Φ(C) is a tower. By the
minimality of F0 , we must have Φ(C) = F0 for every C ∈ Γ. This means
that g(C) ∈ Γ for all C ∈ Γ, so Γ is also a tower; by minimality again,
Γ = F0 . This shows that F0 is totally ordered.
Now let A be the union of all sets in F0 . Then A ∈ F0 by the second tower
property, and g(A) ∈ F0 by the third. But A is the largest member of F0
and A ⊂ g(A), so A = g(A). 

Now let F be the collection of all totally ordered subsets of a partially

ordered set E. Since every single-element subset of E is totally ordered, F
is not empty. Note that the union of any chain of totally ordered sets is
totally ordered. Now let f be a choice function for E. If A ∈ F, let A∗ be
the set of all x in the complement of A such that A ∪ {x} ∈ F. If A∗ 6= ∅,
let
g(A) = A ∪ {f (A∗ )}.
 19

If A∗ = ∅, let g(A) = A.
By the lemma, A∗ = ∅ for at least one A ∈ F, and any such A is a maximal
element of F. 
Problem 7: Consider the curve Γ = {y = f (x)} in R2 , 0 ≤ x ≤ 1. Assume
that f is twice continuously differentiable in 0 ≤ x ≤ 1. Then show that
m(Γ + Γ) > 0 if and only if Γ + Γ contains an open set, if and only if f is
not linear.
Solution. We are asked to show the equivalence of the conditions (i) m(Γ +
Γ) > 0, (ii) Γ + Γ contains an open set, and (iii) f is not linear. We will
show that (ii) implies (i), which implies (iii), which implies (ii).
First, we should note that Γ+Γ is measurable. The problem doesn’t ask for
this, but it’s worth pointing out. Consider G : [0, 1] × [0, 1] → R2 defined
by G(x, y) = (x + y, f (x) + f (y)). Then Γ + Γ is just the range of G. Since
differentiable functions map measurable sets to measurable sets, Γ + Γ is
measurable. (We haven’t proved yet that differentiable functions preserve
measurability, but I assume we will once we get further into differentiation
theory.)
The easiest of our three implications is (ii) implies (i). Suppose Γ + Γ
contains an open set. Open sets have positive measure, so Γ + Γ is a
measurable set with a subset of positive measure, so it has positive measure.
Now suppose m(Γ + Γ) > 0. We wish to show that f is not linear. Suppose
instead that f is linear, say f (x) = ax+b. Then for any x, x0 , (x+x0 , f (x)+
f (x0 )) = (x + x0 , a(x + x0 ) + 2b) so Γ + Γ is a subset of the line y = ax + 2b,
which has measure 0. Thus, if Γ + Γ has positive measure, f must not be
linear.
The third implication is the least trivial. Suppose f is not linear. Then
there are points x0 , y0 ∈ [0, 1] with f 0 (x0 ) 6= f 0 (y0 ). Then the Jacobian
1 1
DG = = f 0 (y) − f 0 (x)
f 0 (x) f 0 (y)
is nonzero at the point (x, y) ∈ [0, 1] × [0, 1], where G(x, y) = (x + y, f (x) +
f (y)) as above. WLOG we may assume (x, y) ∈ (0, 1) × (0, 1) since a
nonlinear function on [0, 1] with continuous derivative cannot have con-
stant derivative everywhere on (0, 1). Then the Inverse Function Theorem
guarantees that there is an open neighborhood of (x, y) on which G is a
diffeomorphism; since diffeomorphisms are homeomorphisms, this implies
that the image of G contains an open set. 

Chapter 2.5, Page 89

Exercise 1: Given a collection of sets F1 , . . . , Fn , construct another collection
Sn SN
F1∗ , . . . , FN∗ , with N = 2n − 1, so that k=1 Fk = j=1 Fj∗ ; the collection
{Fj∗ } is disjoint; and Fk = F ∗ ⊂Fk Fj∗ for every k.
S
j

Solution. For j = 1, . . . , N , create Fj∗ as follows: First write j as an n-digit

binary number j1 j2 . . . jn . Then for k = 1, . . . , n, let
(
Fk jk = 1
Gk =
Fkc jk = 0.
20

Finally, let
n
\
Fj∗ = Gk .
k=1
For example, 2 = 000 . . . 10 in binary, so
F2∗ = F1c ∩ F2c ∩ · · · ∩ Fn−2
c
∩ Fn−1 ∩ Fnc .
Note that the Fj∗ are pairwise disjoint because if j 6= j 0 , then they differ
in some binary digit, say j` 6= j`0 . Suppose WLOG that j` = 1 and j`0 = 0.
Then Fj∗ ⊂ F` whereas Fj∗0 ⊂ F`c , so they are disjoint.
Also, [
Fk = Fj∗ .
Fj∗ ⊂Fk

To see this, note that the RHS is clearly a subset of the LHS since it is a
union of subsets. Conversely, suppose x ∈ Fk . Define x1 , . . . , xn by xi = 1 if
x ∈ Fi and 0 otherwise. Then if m has the binary digits m1 = x1 , . . . , mn =
∗ ∗ ∗
xn , x ∈ Fm by definition of Fm . Since Fm ⊂ Fk , the result follows.
This implies
[n [N
Fi ⊂ Fj∗ .
i=1 j=1
But Fj∗ ⊂
S
Fi for each j, so the reverse inclusion holds as well. 

Exercise 2: In analogy to Proposition 2.5, prove that if f is integrable on Rd

and δ > 0, then f (δx) converges to f (x) in the L1 norm as δ → 1.
Solution. Let  > 0. Since CC (Rd ) is dense in L1 (Rd ), we can choose
g ∈ CC (Rd ) such that kf − gk1 < 3 . We can also choose λ such that
|δ − 1| < λ ⇒ δ1d < 32 . Now let K = supp(g), and choose M such that
K ⊂ {|x| ≤ M }. Let B = {|x| ≤ M + 1}, which is also compact. Because
g is uniformly continuous on B, ∃λ0 > 0 such that λ0 < 1 and for x, y ∈ B,
|x − y| < λ0 ⇒ |g(x) − g(y)| < 6m(B)

. Now suppose we choose δ such that
0
|δ − 1| < min(λ, Mλ+1 ). Then
Z
kf (δx) − f (x)k = |f (δx) − f (x)|
Z Z Z
≤ |f (δx) − g(δx)| + |f (δx) − g(x)| + |g(x) − f (x)|

by the triangle inequality. The third integral is just kf − gk ≤ 3 . By the

dilation property of the integral, the first integral is δ1d kf − gk ≤ 23 3 = 2 .
Now for the second integral. I claim that the integrand is 0 outside B and

at most 6m(B) inside. If x ∈
/ B, then |x| > M + 1 so
λ0 λ0
   
|δx| = |δ||x| ≥ 1 − |x| > 1 − (M + 1) > M,
M +1 M +1
which implies that g(δx) = g(x) = 0. Now suppose x ∈ B. If g(x) 6= 0,
then x ∈ K and
λ0 λ0
   
|δx| ≤ 1 + |x| < 1 + M <M +1
M +1 M +1
 21

so δx ∈ B and |δx − x| < λ0 ⇒ |g(δx) − g(x)| < 6m(B) 

. If g(δx) 6= 0,

then δx ∈ K ⊂ B and again |g(δx) − g(x)| < 6m(B) because of uniform

continuity. Hence the integrand is 0 outside B and is at most 6m(B) inside
 
B, so its integral is at most 6m(B) m(B) = 6 . Putting the pieces together,
we have kf (δx) − f (x)k < . 
Exercise 4: Suppose f is integrable on [0, b], and
Z b
f (t)
g(x) = dt for 0 < x ≤ b.
x t
Prove that g is integrable on [0, b] and
Z b Z b
g(x)dx = f (t)dt.
0 0

Solution. We may assume WLOG that f (t) ≥ 0, since otherwise we can

analyze f + and f − separately. Now let
f (t)
h(x, t) = χ{0<x≤t≤b} .
t
Then h ≥ 0 and h is clearly measurable since it is a quotient of measurable
functions times another measurable function. By Tonelli’s theorem,
Z ∞ Z b
f (t)
h(x, t)dt = χ{0<x≤b} dt
−∞ x t
is a measurable function of x. Note that this is just equal to g(x) for
0 < x ≤ b and 0 elsewhere. Hence g is measurable (in general, g : (0, b] → R
is measurable iff g · χ(0,b] : R → R is). Moreover, Tonelli’s theorem also tells
us that
Z b Z b Z t 
f (t)
g(x) dx = dt dx
0 t
Z0 x

= h(x, t)
R×R
Z b Z t 
= h(x, t) dx dt
0 0
Z b Z t 
f (t)
= dx dt
0 0 t
Z b
f (t)
= t dt
0 t
Z b
= f (t) dt.
0
Note that the fact that this integral is finite implies that g is integrable and
not just measurable. 
Exercise 5: Suppose F is a closed set in R, whose complement has finite
measure, and let δ(x) denote the distance from x to F , that is,
δ(x) = d(x, F ) = inf{|x − y| : y ∈ F }.
22

Consider
Z
δ(y)
I(x) = dy.
R |x − y|2

(a) Prove that δ is continuous by showing that it satisfies the Lipschitz

condition

|δ(x) − δ(y)| ≤ |x − y|.

(b) Show that I(x) = ∞ for each x ∈ / F.

(c) Show that I(x) < ∞ for a.e. x ∈ F . This may be surprising in view of
the fact that the Lipschitz condition cancels only one power of |x − y|
in the integrand of I.

Solution.
(a) Let  > 0. Choose z ∈ F such that |y − z| < δ(y) + . Then

δ(x) ≤ |x − z| ≤ |x − y| + |y − z| < |x − y| + δ(y) +  ⇒ δ(x) − δ(y) < |x − y| + .

Interchanging the roles of x and y, we also have

δ(y) − δ(x) < |x − y| + .

Hence |δ(x) − δ(y)| < |x − y| +  for any  > 0. This implies |δ(x) −
δ(y)| ≤ |x − y|.
(b) Suppose x ∈ / F . Because F is closed, this implies δ(x) > 0, since
otherwise there would be a sequence of points in F converging to x.
Let λ = δ(x). By the Lipschitz condition from part (a), |x − y| < λ2 ⇒
|δ(y) − λ| < λ2 ⇒ δ(y) ≥ λ2 . Hence

Z ∞
δ(y)
I(x) = dy
−∞ |x − y|2
Z x+λ/2
δ(y)
≥ dy
x−λ/2 |x − y|2
Z x+λ/2 λ
2
≥ dy
x−λ/2 |x − y|2
Z λ/2
λ 1
= dy = ∞.
2 −λ/2 y2

(c) First, consider that for y ∈

/ F,
Z Z ∞
1 1 2
dx ≤ 2 dx =
F |x − y|2 δ(y) x2 δ(y)
 23

since F ⊂ {x : |x − y| ≥ δ(y)}. Now since I(x) ≥ 0, we have by

Tonelli’s theorem
Z Z
δ(y)
I(x) dx = χ (x)
2 F
F R×R |x − y|
Z Z 
δ(y)
= 2
χF (x) dx dy
R R |x − y|
Z Z 
1
= δ(y) 2
dx dy
c F |x − y|
ZF
2
≤ δ(y) dy
Fc δ(y)
= 2m(F c ) < ∞.
R
Since F I(x) dx < ∞, we must have I(x) < ∞ for almost all x ∈ F .
(This is actually not all that shocking, since I(x) is clearly less than ∞
for an interior point. Of course, there are closed sets whose boundaries
have positive measure, but those are the nasty guys.)


Exercise 6: Integrability of f on R does not necessarily imply the conver-

gence of f (x) to 0 as x → ∞.
(a) There exists a positive continuous function f on R such that f is
integrable on R, yet lim supx→∞ f (x) = ∞.
(b) However, if we assume that f is uniformly continuous on R and inte-
grable, then lim|x|→∞ f (x) = 0.

Solution.
(a) Let
(  1
c
1


23n+4 d x, n, n + 22n+1 n≤x≤n+ 22n+1 , n ∈Z
f (x) =
0 else.

The graph of f consists of a series of triangular spikes withR height

2n+2 and base 22n+1 1
. The nth such spike has area 2−n , so |f | =
P ∞ −n
n=0 2 = 2. But, because the spikes get arbitrarily high, lim sup f (x) =
∞.
(b) Suppose f is uniformly continuous on R, and let  > 0. Select δ > 0
such that |x − y| < δ ⇒ |f (x) − f (y)| < 2 , and also require δ < 12 .
Since f (x) 6→ 0, ∃x1 > 0 such that |f (x1 )| ≥ . Then |f (y)| ≥ 2 for
y ∈ (x1 −δ, x1 +δ). Now since f (x) 6→ 0, ∃x2 > x1 +1 with |f (x2 )| ≥ .
Then |f | ≥ 2 on (x2 − δ, x2 + δ). Continuing in this manner, we obtain
infinitely many intervals of length 2δ on which |f | ≥ 2 . These intervals
are disjoint because of our requirements that |xn+1 −x| > 1 and δ < 12 .
Hence, by Tchebycheff’s inequality,
Z
 
|f (x)| dx ≥ m({x : |f (x)| ≥ }) = ∞.
R 2 2

24

Exercise 8: If f is integrable on R, show that

Z x
F (x) = f (t)dt
−∞

is uniformly continuous.

Solution. Let  > 0. By the absolute

R continuity of the integral (Prop 1.12b),
∃δ > 0 such that m(E) < δ ⇒ E |f | < . Then (assuming WLOG x > y),
Z x Z x
|x − y| < δ ⇒ |F (x) − F (y)| = f (t)dt ≤ |f (t)| dt < 
y y

because m([y, x]) = |x − y| < δ. 

Exercise 10: Suppose f ≥ 0, and let E2k = {x : f (x) > 2k } and Fk = {x :

2k < f (x) ≤ 2k+1 }. If f is finite almost everywhere, then
∞
[
Fk = {f (x) > 0}
k=−∞

and the sets Fk are disjoint.

Prove that f is integrable if and only if
∞
X ∞
X
2k m(Fk ) < ∞, if and only if 2k m(E2k ) < ∞.
k=−∞ k=−∞

Use this result to verify the following assertions. Let

(
|x|−a if |x| ≤ 1
f (x) =
0 otherwise
and (
|x|−b if |x| > 1
g(x) =
0 otherwise.
Then f is integrable on Rd if and only if a < d; also g is integrable on Rd
if and only if b > d.

Solution. Let
∞
X
g(x) = 2k χFk (x),
k=−∞
∞
X
h(x) = 2k+1 χFk (x).
k=−∞
Then g(x) ≤ f (x) ≤ h(x) by definition of Fk . Then
Z Z ∞
X
f (x)dx < ∞ ⇒ g(x)dx = 2k m(Fk ) < ∞
k=−∞

whereas
∞
X Z Z ∞
X ∞
X
k k+1
2 m(Fk ) < ∞ ⇒ f (x)dx < h(x)dx = 2 m(Fk ) = 2 2k m(Fk ) < ∞.
k=−∞ k=−∞ k=−∞
 25

Now let
∞
X
φ(x) = 2k χEk (x).
k=−∞
Pk
Then f (x) ≤ φ(x) ≤ 2f (x) because if 2k < f (x) ≤ 2k+1 , φ(x) = j=−∞ 2k =
1 + 1 + 2 + 4 + · · · + 2k = 2k+1 . Hence
Z Z ∞
X
f (x)dx < ∞ ⇔ φ(x)dx = 2k m(Ek ) < ∞.
k=−∞

Now for the function f given,

(
k {|x| ≤ 1} k≤0
Ek = {f (x) > 2 } =
{|x| ≤ 2−k/a } k≥1
so (
2d k≤0
m(Ek ) = .
2d 2−kd/a k≥1
So f is integrable iff
∞
X 0
X ∞
X ∞
X
2k m(Ek ) = 2k 2d + 2k 2d 2−kd/a = 2d+1 + 2d 2(1−d/a)k < ∞.
k=−∞ k=−∞ k=1 k=−∞

This infinite sum will converge iff the constant 1 − ad is negative, i.e. iff
a < d.
For the function g given, let us redefine g(x) = 1 for |x| ≤ 1; clearly this
does not affect the integrablity of g. Now Ek is empty for k > 0, so we
need only consider negative values of k.
g(x) > 2k ⇔ |x| < 2−k/b
so Ek is a cube of volume 2d 2−kd/b . Hence g is integrable iff
0
X 0
X
2k 2d 2−kd/b = 2d 2(1−d/b)k
k=−∞ k=−∞

converges. This will happen iff 1 − d/b > 0 ⇒ b > d. 

R
Exercise 11: Prove that if f is integrable on Rd , and R E f (x)dx ≥ 0 for every
measurable E, then f (x) ≥ 0 a.e. x. As a result, if E f (x)dx = 0 for every
measurable E, then f (x) = 0 a.e.
Proof. Suppose it is not true that f (x) ≥ 0 a.e., so m({f (x) < 0} is positive.
Now
∞   ∞  
[ 1 X 1
{x : f (x) < 0} = x : f (x) < − ⇒ m({x : f (x) < 0}) ≤ m x : f (x) < −
n=1
n n=1
n
by countable additivity. Hence at least one of the sets
 
1
En = x : f (x) < −
n
has positive measure. But then
Z Z
1 1
f (x)dx ≤ − dx = − m(En ) < 0.
En En n n
26

R
By contraposition, if E f (x)dx ≥ 0 for every measurable set E, then f (x) ≥
0 a.e. R R
RNow if E f (x)dx = 0 for every measurable E, then E f (x)dx ≥ 0 and
E
−f (x)dx ≥ 0, which means f ≥ 0 a.e. and −f ≥ 0 a.e. Hence f = 0
a.e. 
Exercise 12: Show that there are f ∈ L1 (Rd ) and a sequence {fn } with
fn ∈ L1 (Rd ) such that
kf − fn k1 → 0,
but fn (x) → f (x) for no x.
Solution. To assist in constructing such a sequence, we first construct a
sequence of measurable sets En ⊂ Rd with the property that m(En ) → 0
but every x ∈ Rd is in infinitely many En . We proceed as follows: Choose
integers N1 , N2 , . . . such that
1 1
1 + + ··· + > 10
2 N1
1 1 1
+ + ··· + > 100
N1 + 1 N1 + 2 N2
1 1 1
+ + ··· + > 1000
N2 + 1 N2 + 2 N3
etc. This is possible because of the divergence of the harmonic series.
For convience, we also define N0 = 0. Next, for each k = 0, 1, 2, . . . , let
BNk +1 be the cube of volume Nk1+1 centered at the origin. Then, for a
given k, define Bj for Nk + 1 < j ≤ Nk+1 to be the cube centered at the
origin with |Bj | = |Bj−1 | + 1j . Finally, we define ENk +1 = BNk +1 , and for
Nk + 1 < j ≤ Nk+1 , Ej = Bj \ Bj−1 . I claim that the sets En have the
desired properties. First, note that m(En ) = n1 . This is obvious for Nk + 1;
for Nk +1 < j ≤ Nk+1 it is easy to see inductively that |Bj | = Nk1+1 +· · ·+ 1j
and since they are nested sets, |Bj \ Bj−1 | 1j . Thus m(En ) → 0. However,
for each k,
Nk+1
[
Ej
j=Nk +1

is a cube centered at the origin with a volume greater than 10k . For any
given x, these cubes will eventually contain x, i.e. there is some K such
that
Nk+1
[
k>K⇒x∈ Ek .
j=Nk +1
Hence every x is in infinitely many Ej as desired.
Having constructed these sets, we simply let fn (x) = χEn (x) and f (x) = 0.
L1
Then fn (x)dx = n1 so kfn − f k = |fn − f | = fn → 0, i.e. fn → f .
R R R

However, for any given x there are infinitely many n such that fn (x) = 1,
so fn (x) 6→ f (x) for any x. 
Exercise 13: Give an example of two measurable sets A and B such that
A + B is not measurable.
 27

Solution. As suggested in the hint, let N ⊂ R be a non-measurable set.

Let A = {0} × [0, 1] and B = N × {0}. Then A and B are both measurable
subsets of R2 because they are subsets of lines, which have measure 0. Now
A + B = N × [0, 1]. By Proposition 3.4, if N × [0, 1] were measurable, since
[0, 1] has positive measure, this would imply that N was measurable too, a
contradiction. Hence A + B is not measurable. 
Exercise 15: Consider the function defined over R by
(
x−1/2 if 0 < x < 1,
f (x) =
0 otherwise.
For a fixed enumeration {rn } of the rationals Q, let
∞
X
F (x) = 2−n f (x − rn ).
n=1
Prove that F is integrable, hence the series defining F converges for almost
every x ∈ R. However, observe that this series is unbounded on every
interval, and in fact, any function F̃ that agrees with F a.e. is unbounded
in any interval.
Solution. First we compute the integral of f ; the improper Riemann inte-
gral is
Z 1
1 √ 1
√ dx = 2 x = 2,
0 x 0
but we only proved that the Lebesgue and Riemann integrals are equal for
the proper Riemann integral. Of course it’s true for improper integrals as
well; here, since (0, 1] = ∪∞ 1 1
n=1 ( n+1 , n ], we have by countable additivity that
Z X∞ Z
f dx = f dx
1 1
R n=1 ( n+1 ,n ]
N Z
X
= lim f dx
N →∞ 1
( n+1 1
,n ]
n=1
Z 1
= lim f dx
N →∞ 1
N +1
Z 1
= lim f dx (since this limit exists)
a→0 a
= 2.
By translation invariance, the integral of f (x − rn ) is also 2. Now since f
is nonnegative everywhere, the partial sums are monotonely increasing, so
by the Monotone Convergence Theorem
Z ∞ Z
X X∞
−n
F dx = 2 f (x − rn )dx = 21−n = 2.
n=1 n=1
Since this integral is finite, F is integrable. This implies that F is finite-
valued for almost all x ∈ R.
Now let F̄ be any function that agrees with F almost everywhere, and I
28

any interval on the real line. Let rN be some rational number contained in
I. Then for any M > 0, f (x−rN ) > M on the interval (rN − M12 , rN + M12 ),
which intersects I in an interval IM of positive measure. Since F̄ agrees
with F almost everywhere, it must also be greater than M at almost all
points of this interval IM ⊂ I. Hence F̄ exceeds any finite value M on
I. 
Exercise 17: Suppose f is defined on R2 as follows: f (x, y) = an if n ≤ x <
n + 1 and n ≤ y < n + 1, n ≥ 0; f (x, y) = −an if n ≤ xP< n + 1 and
n + 1 ≤ y < n + 2, n ≥ 0; f (x, y) = 0 elsewhere. Here an = k≤n bk , with
P∞
{bk } a positive sequence such that k=0 bk = s < ∞. R
y
(a) Verify that each
R Rslice f and fx is integrable. Also for all x, fx (y)dy =
0, and hence
R ( f (x, y)dy)dx = 0. R y
(b) However, f y (x)dx = a0 if 0 ≤ y < 1, and R y f (x)dx = an − an−1
if n ≤ y < n + 1 with n ≥ 1. Hence y 7→ f (x)dx is integrable on
(0, ∞) and
Z Z 
f (x, y)dx dy = s.
R
(c) Note that R×R |f (x, y)|dxdy = ∞.

Solution.
(a) Since f is constant on boxes and 0 elsewhere, the horizontal and ver-
tical slices are constant on intervals and 0 elsewhere, and therefore
integrable. More precisely,

−abyc−1 byc − 1 ≤ x < byc

y
f (x) = abyc byc ≤ x < byc + 1

0 else


for y ≥ 1, (
y a0 0≤x<1
f (x) =
0 else
for 0 ≤ y < 1, and

abxc
 bxc ≤ y < bxc + 1
fx (y) = −abxc bxc + 1 ≤ y < bxc + 2

0 else


where bxc is the greatest integer less than or equal to x. (For x < 0
the function fx (y) is identically
R 0, and for y < 0 the function f y (x)
is identically 0.) Clearly fx (y)dy = 0 for all x, since fx (y) is equal
to abxc on an interval of Rlength
R 1 and −abxc on an interval of length 1
and 0 elsewhere. Hence f (x, y)dydx = 0.
(b) Since all the integrals are of constants on intervals Rof length 1, it
immediately follows from the formulas in part (a) that f y (x)dx is a0
for 0 ≤ y < 1 and an − an−1 = bn for n ≤ y < n + 1. Then
Z Z  X ∞ Z n+1 Z  ∞
X
y y
f (x)dx = f (x)dx dy = bn = s.
R R n=0 n R n=0
 29

(c) Since |f (x, y)| is positive, we may use Tonelli’s theorem, so

Z Z Z 
|f (x, y)| = fx (y)dy dx
R×R
∞ Z
X n+1 Z 
= fx (y)dx dx
n=0 n

X∞
= 2an = ∞
n=0

since an > a0 so the terms in the sum are bounded away from 0.


Exercise 18: Let f be a measurable finite-valued function on [0, 1], and sup-
pose that |f (x) − f (y)| is integrable on [0, 1] × [0, 1]. Show that f (x) is
integrable on [0, 1].

Solution. Let g(x, y) = |f (x) − f (y)|. By Fubini’s Theorem, since g is

integrable on [0, 1] × [0, 1], g y (x) is an integrable function of x for almost
all y ∈ [0, 1]. Choose any such y. Then since f (x) − f (y) ≤ |f (x) − f (y)|,
Z 1 Z 1
(f (x) − f (y)) dx ≤ |f (x) − f (y)|dx < ∞
0 0
so
Z 1 Z 1
f (x)dx ≤ f (y) + |f (x) − f (y)|dx < ∞.
0 0


Exercise 19: Suppose f is integrable on Rd . For each α > 0, let Eα = {x :

|f (x)| > α}. Prove that
Z Z ∞
|f (x)|dx = m(Eα )dα.
Rd 0

Solution. By Tonelli’s Theorem,

Z ∞ Z ∞ Z 
m(Eα )dα = χ|f (x)|>α dx dα
0 0 Rd
Z Z ∞ 
= χ|f (x)|>α dα dx
d 0
ZR
= |f (x)|dx.
Rd

Exercise 22: Prove that if f ∈ L1 (Rd ) and

Z
fˆ(ξ) = f (x)e−2πixξ dx,
Rd

then fˆ(ξ) → 0 as |ξ| → ∞. (This is the Riemann-Lebesgue lemma.)

30

Solution. By translation invariance,

Z
fˆ(ξ) = f (x)e−2πix·ξ dx
d
ZR  
ξ ξ
−2πi(x− 2|ξ| 2 )·ξ
= f x− 2
e dx
Rd 2|ξ|
Z  
ξ 2πi ξ ·ξ
= f x − 2 e−2πix·ξ e 2|ξ|2 dx
Rd |ξ|
Z  
ξ
= −f x − e−2πix·ξ dx
Rd 2|ξ|2
so, multiplying by 12 and adding the original expression,
Z Z   
1 ξ
fˆ(ξ) = f (x)e−2πix·ξ dx − f x− e −2πix·ξ
dx
2 Rd Rd 2|ξ|2
Z   
1 ξ
= f (x) − f x − e−2πix·ξ dx
2 Rd 2|ξ|2
and
Z   
ˆ 1 ξ
|f (ξ)| = f (x) − f x − e−2πix·ξ dx
2 Rd 2|ξ|2
Z   
1 ξ
≤ f (x) − f x − e−2πix·ξ dx
2 Rd 2|ξ|2
Z  
1 ξ
= f (x) − f x − dx
2 Rd 2|ξ|2
 
1 ξ
= f (x) − f x − .
2 2|ξ|2 1
ξ ξ 1
As |ξ| → ∞, | 2|ξ|2 | → 0, so kf (x) − f (x − 2|ξ|2 )k1 → 0 by the L -continuity

of translation (Proposition 2.5). 

Exercise 23: As an application of the Fourier transform, show that there
does not exist a function I ∈ L1 (Rd ) such that
f ∗I =f for all f ∈ L1 (Rd ).

Solution. Suppose such an I exists. Then for every f ∈ L1 , fˆ(ξ)I(ξ)

ˆ = fˆ(ξ)
ˆ
for all ξ. This implies that I(ξ) = 1 for all ξ. But this contradicts the
Riemann-Lebesgue Lemma (Problem 22). QED.
Note: To be totally complete, we should show that for any ξ there is a
function g ∈ L1 such that ĝ(ξ) 6= 0. Otherwise, the equation fˆIˆ = fˆ
wouldn’t necessarily imply Iˆ = 1 everywhere. But it is easy to show that
such a g exists for any ξ; for example, g could be equal to e2πix·ξ on some
compact set and 0 outside. 
Exercise 24: Consider the convolution
Z
(f ∗ g)(x) = f (x − y)g(y)dy.
Rd
(a) Show that f ∗ g is uniformly continuous when f is integrable and g
bounded.
 31

(b) If in addition g is integrable, prove that (f ∗ g)(x) → 0 as |x| → ∞.

Solution.
(a) Since g is bounded, ∃M with |g| < M everywhere. Then
Z Z
|f ∗ g(x) − f ∗ g(x0 )| = f (x − y)g(y)dy − f (x0 − y)g(y)dy
Rd Rd
Z
= (f (x − y) − f (x0 − y)) g(y)dy
Rd
Z
≤ |f (x − y) − f (x0 − y)| |g(y)|dy
Rd
Z
≤M |f (x − y) − f (x0 − y)|dy
d
ZR
=M |f (y) − f (y + (x0 − x))|dy
Rd
= M kf (y) − f (y + (x − x0 ))k1 .
In the penultimate
R step
R we have used translation invariance and the
fact that f (y)dy = f (−y)dy provided both integrals are taken over
all of Rd . Now by the L1 -continuity of translation, ∃δ > 0 such that
|x − x0 | < δ ⇒ kf (y) − f (y + (x − x0 ))k1 < M 
. This in turn implies
0
|f ∗ g(x) − f ∗ g(x )| < , so f ∗ g is uniformly continuous.
(b) Let  > 0. Since CC (Rd ) is dense in L1 (Rd ), we may choose f˜ such
that supp(f˜) = K is compact, f˜ is continuous, and kf − f˜k1 < 2M 
,
where M is a bound for |g| as in part (a). Continuous functions with
compact support are bounded, so choose N such that |f˜| < N . Now
Z
|f ∗ g(x)| = f (x − y)g(y)dy
Rd
Z  
= f˜(x − y) + (f (x − y) − f˜(x − y)) g(y)dy
d
ZR
≤ f˜(x − y) + (f (x − y) − f˜(x − y))g(y) dy
Rd
Z Z
≤ |f˜(x − y)||g(y)|dy + f (x − y) − f˜(x − y) |g(y)|dy.
Rd Rd
Call the first integral I1 and the second I2 . Since |g| < M , I2 ≤
M kf − f˜k1 <R 2 . Now since g is integrable, there must exist compact

F such that F c |g| < 2N . Then if |x| is larger than the sum of the
diameters of K and N ,
Z
I2 = |f˜(x − y)||g(y)|dy
Rd
Z
= |f˜(x − y)||g(y)|dy since f˜ = 0 on K c
x−y∈K
Z
≤N |g(y)|dy
x−y∈K
Z

≤N |g(y)|dy <
F c 2
32

since {y : x − y ∈ K} ⊂ F c . Thus, for sufficiently large x, |f ∗ g(x)| ≤

I1 + I2 < .


Chapter 2.6, Page 95

R 2π
Problem 1: If f is integrable on [0, 2π], then 0 f (x)e−inx dx → 0 as |n| →
∞. Show as a consequence that if E is a measurable subset of [0, 2π], then
Z
m(E)
cos2 (nx + un )dx → , as n → ∞
E 2
for any sequence {un }.
R 2π R 2π
Solution. First, note that 0 f (x) cos(nx)dx → 0 and 0 f (x) sin(nx) →
R 2π
0 since these are the real and imaginary parts of 0 f (x)e−inx dx. In
particular, if we let f (x) = χE (x) for some measurable E ⊂ [0, 2π], then for
R 2π R 2π
any  > 0, ∃N such that | 0 χE (x) sin(nx)dx| and | 0 χE (x) cos(nx)dx|
are both less than 2 provided |n| > N . Then for any sequence un ,
Z Z
cos(2nx + 2un )dx = cos(2nx) cos(2un ) − sin(2nx) sin(2un )dx
E E
Z 2π Z 2π
= cos(2un ) χE (x) cos(2nx)dx − sin(2un ) χE (x) sin(2nx)dx
0 0
Z 2π Z 2π
≤ | cos(2un )| χE (x) cos(2nx)dx + | sin(2un )| χE (x) sin(2nx)dx
0 0
 
≤1· +1· =
2 2
R
for |n| > N . Hence E cos(2nx + un )dx → 0 as |n| → ∞. Now
Z Z
1
cos2 (nx + un ) = (1 + cos(2(nx + un ))) dx
E E 2
Z 2π
m(E)
= + χE (x) cos(2nx + 2un )dx
2 0

and we have shown that the second term tends to 0 as |n| → ∞. 

Problem 2: Prove the Cantor-Lebesgue theorem: if

∞
X ∞
X
An (x) = (an cos nx + bn sin nx)
n=0 n=0

converges for x in a set of positive measure (or in particular for all x), then
an → 0 and bn → 0 as n → ∞.
p
Solution. We can rewrite An (x) = cn cos(nx + dn ) where cn = a2n + b2n
P dn is some phase angle (it can be − arctan(bn /an ), for example). If
and
An (x) converges on some E with m(E) > 0, then An (x) → 0 on E. By
 33

Egorov’s theorem, this implies An → 0 uniformly on some E 0 ⊂ e with

m(E 0 ) > 0. Then
u
cn cos(nx + dn ) → 0 on E 0
u
⇒ c2n cos2 (nx + dn ) → 0 on E 0
Z
⇒ c2n cos2 (nx + dn )dx → 0.
E0
R0 0
But E cos2 (nx+dn )dx → m(E 2
)
by the previous problem, so c2n → 0, which
implies cn → 0, which implies an → 0 and bn → 0. 

Problem 3: A sequence {fk } of measurable functions on Rd is Cauchy in

measure if for every  > 0,

m({x : |fk (x) − f` (x)| > }) → 0 as k, ` → ∞.

We say that {fk } converges in measure to a (measurable) function f if

for every  > 0,

m({x : |fk (x) − f (x)| > }) → 0 as k → ∞.

This notion coincides with the “convergence in probability” of probability

theory.
Prove that if a sequence {fk } of integrable functions converges to f in L1 ,
then {fk } converges to f in measure. Is the converse true?

Solution. Suppose fn , f ∈ L1 and fn → f in L1 . By Chebyshev’s Inequal-

ity,
kfn − f k1
m ({x : |fn (x) − f (x)| > }) ≤

and since the RHS tends to 0 as n → ∞, the LHS does as well, so fn →
f in measure. However, the converse does not hold. Consider the case
f (x) = 0, fn (x) = nχ[0, n1 ] . Then m({x : f (x) 6= fn (x)}) = n1 so for any ,
m({x : |f (x) − fn (x)| > }) ≤ n1 → 0. Hence fn → f in measure. However,
kfn − f k1 = 1 for all n, so fn 6→ f in L1 . 

Problem 4: We have already seen (in Exercise 8, Chapter 1) that if E is a

measurable set in Rd , and L is a linear transformation of Rd to Rd , then
L(E) is also measurable, and if E has measure 0, then so has L(E). The
quantitative statement is

m(L(E)) = | det(L)|m(E).

As a special case, note that the Lebesgue measure is invariant under rota-
tions. (For this special case see also Exercise 26 in the next chapter.)
The above identity can be proved using Fubini’s theorem as follows.
(a) Consider first the case d = 2, and L a “strictly” upper triangular
transformation x0 = x + ay, y 0 = y. Then

χL(E) (x, y) = χE (L−1 (x, y)) = χE (x − ay, y).

34

Hence
Z Z 
m(L(E)) = χE (x − ay, y)dx dy
R×R
Z Z 
= χE (x, y)dx dy
R×R
= m(E),
by the translation-invariance of the measure.
(b) Similarly m(L(E)) = m(E) if L is strictly lower triangular. In general,
one can write L = L1 ∆L2 , where Lj are strictly (upper and lower)
triangular and ∆ is diagonal. Thus m(L(E)) = | det L|m(E), if one
uses Exercise 7 in Chapter 1.
Solution.
(a) I’m not quite sure what to comment on here, since the problem state-
ment pretty much did all the work for me. I guess I should point out
that the use of Tonelli’s theorem to turn the double integral into an
iterated integral is justified because χE is nonnegative; note that this
proves that the result holds even if m(E) = ∞.
(b) The fact that lower triangular transformations work the same way is
obvious. Now supposing L = L1 ∆L2 , we have
m(L(E)) = m(L1 (∆(L2 (E)))) = m(∆(L2 (E)))
= | det(∆)|m(L2 (E)) = | det(∆)|m(E) = | det(L)|m(E)
since det(L) = det(L1 ) det(∆) det(L2 ) = det(∆) and m(∆(E)) =
| det(∆)|m(E) by Exercise 7 of Chapter 1. Thus, every linear transfor-
mation that has an LU decomposition works as we want it to. How-
ever, I think the problem is flawed because not every matrix has an
LU decomposition (in fact, not even every invertible matrix does.) In
particular, in the 2 × 2 case a matrix of the form L1 ∆L2 will look like
either
     
1 c d1 0 1 0 d1 + cd2 e cd2
=
0 1 0 d2 e 1 d2 e d2
or
     
1 0 d1 0 1 c d1 cd1
= .
e 1 0 d2 0 1 ed1 ed1 c + d2
But a matrix of the form
 
0 a
b 0
with a, b 6= 0 cannot be put in either form, because in the first case we
would need d2 = 0 in order to make the lower right entry 0, and then
the upper right and lower left entries could not be nonzero; similarly,
in the second case, we would need d1 = 0 which would make a, b 6= 0
impossible. Hence, there are some matrices that cannot be factored
in the way this problem indicates; for the ones that can, though, we
know that they expand measures by a factor of | det |.

 35

Chapter 3.5, Page 145

R
Exercise 1: Suppose φ is an integrable function on Rd with Rd φ(x)dx = 1.
Set Kδ (x) = δ −d φ(x/δ), δ > 0.
(a) Prove that {Kδ }δ>0 is a family of good kernels.
(b) Assume in addition that φ is bounded and supported in a bounded
set. Verify that {Kδ }δ>0 is an approximation to the identity.
(c) Show that Theorem 2.3 holds for good kernels as well.
Solution.
(a) By
R the dilation Rproperties of the integral, R we have immediately
R that
Rd
K δ (x)dx = Rd
φ(x)dx = 1 and Rd
|Kδ (x)|dx = Rd
|φ(x)|dx =
kφk1 < ∞. This proves the first two properties of good kernels. For
1
the last, we recall that for R φ ∈ L , for every  > 0 there exists a
compact set F such that F c |φ| < . Now compact subsets of Rd are

bounded, so K ⊂ Br (0) for some radius Rr . Now if h > 0 is any fixed
number, for δ < rh this will imply that |x|>h |Kδ (x)|dx < . Thus,
R
for any h > 0, |x|>h |Kδ (x)|dx → 0 as δ → 0. Hence {Kδ } is a family
of good kernels.
(b) Suppose |φ| ≤ M everywhere and φ(x) = 0 for |x| ≥ B. Let A =
M B d+1 . Then for any δ > 0,
1 x
|Kδ (x)| = δ −(d+1) φ ≤ M δ −(d+1)
δ δ
A A
≤ ≤
(δB)d+1 |x|d+1
for xδ ≤ B; for xδ > B we have Kδ (x) = 0. Hence |Kδ (x)| ≤ Aδ/|x|d+1
for all x and δ, so {Kδ } is an approximation to the identity.
(c) Suppose {Kδ } is any family of good kernels. Then
Z
kf ∗ Kδ − f k = |f ∗ Kδ (x) − f (x)| dx
d
ZR Z
= f (x − y)Kδ (y)dy − f (x) dx
Rd Rd
Z Z
= (f (x − y) − f (x))Kδ (y)dy dx
d d
ZR Z R
≤ |f (x − y) − f (x)||Kδ (y)|dydx
Rd Rd
Z Z
= |f (x − y) − f (x)||Kδ (y)|dxdy
d d
ZR RZ Z Z
= |f (x − y) − f (x)||Kδ (y)|dxdy + |f (x − y) − f (x)||Kδ (y)|dxdy
|y|≤η Rd |y|>η Rd

for any η > 0. Let us call the first integral I1 and the second I2 . Then
Z Z 
I1 = |Kδ (y)| |f (x − y) − f (x)|dx dy
|y|≤η Rd
Z
= |Kδ (y)|kf (x − y) − f (x)k1 dy.
|y|≤η
36

Now by the L1 -continuity of translation, kf (x − y) − f (x)k1 → 0 as

y → 0. Thus, if η is sufficiently small, this norm will be at most, say,

2A for all y ∈ [−η, η]. Then
Z
 
I1 ≤ |Kδ (y)| dy ≤
|y|≤η 2A 2
R
since Rd |Kδ (y)|dy ≤ A for all δ. Thus, by choosing η sufficiently
small, we may make I1 as small as we like, independent of δ. On the
other hand,
Z Z
I2 = |Kδ (y)| |f (x − y) − f (x)|dxdy
|y|>η Rd
Z Z
≤ |Kδ (y)| (|f (x − y)| + |f (x)|)dxdy
|y|>η Rd
Z
= |Kδ (y)| · 2kf k1 dy
|y|>η
Z
= 2kf k1 |Kδ (y)|dy → 0
|y|>η

as δ → 0. Putting the two halves together, we see that choosing a

sufficiently small η and then letting δ → 0 makes kf ∗ Kδ − f k1 → 0.


Exercise 3: Suppose 0 is a point of (Lebesgue) density of the set E ⊂ R.

Show that for each of the individual conditions below there is an infinite
sequence of points xn ∈ E, with xn 6= 0, and xn → 0 as n → ∞.
(a) The sequence also satisfies −xn ∈ E for all n.
(b) In addition, 2xn belongs to E for all n.
Generalize.

Solution.
(a) Since 0 is a point of density of E, there exists r0 > 0 such that m(E ∩
Br (0)) > 32 m(Br (0)) = 34 r for r ≤ r0 . By symmetry, m((−E) ∩
Br0 (0)) = m(E∩Br0 (0)) ≥ 34 r0 . Now E and −E both intersect Br0 (0),
which has measure 2r0 , in sets of measure at least 34 r0 . Therefore they
intersect each other in a set of measure at least 23 r0 . Since E ∩ (−E)
has positive measure, it is infinite, so it contains an infinite sequence
xn . This sequence satisfies xn ∈ E and −xn ∈ E.
(b) Since 0 is a point of density of E, there exists r0 > 0 such that
m(E ∩ Br (0)) > 23 m(Br (0)) = 43 r for r ≤ r0 . Let E0 = E ∩ Br0 (0).
Then m( 21 E0 ) = 12 m(E0 ) ≥ 23 r0 as we showed in a previous homework
about the effect of dilation on Lebesgue measure. Now 21 E0 = ( 12 E) ∩
Br0 /2 (0) has measure at least 23 r0 , and we also know m(E∩Br0 /2 (0)) ≥
2 2 r0 1
3 m(Br0 /2 ) = 3 r0 since 2 < r0 . So E and 2 E both intersect Br0 /2 ,
which has measure r0 , in sets of measure at least 23 r0 . Therefore they
intersect each other in a set of measure at least 13 r0 . Since E ∩ ( 12 E)
has positive measure, it must contain an infinite sequence xn . Then
xn ∈ E and 2xn ∈ E.
 37

Clearly the above process generalizes to produce a sequence xn with xn ∈ E

and cxn ∈ E for any c 6= 0. 

Exercise 5: Consider the function on R defined by

(
1
2 if |x| ≤ 1/2,
f (x) = |x|(log 1/|x|)
0 otherwise.
(a) Verify that f is integrable.
(b) Establish the inequality
c
f ∗ (x) ≥ for some c > 0 and all |x| ≤ 1/2,
|x|(log 1/|x|)
to conclude that the maximal function f ∗ is not locally integrable.

Solution.
(a) We have
1/2
Z Z 1/2
1 1 2
|f (x)|dx = 2 dx = 2 = .
Rd 0 x(log 1/x)2 log x1 log 2
0
1
(b) For 0 < |x| ≤ if B = (0, 2x) is the ball of radius |x| centered at x,
2,
then
Z Z 2
1 1
|f (y)|dy = |x||f (y)|dy
m(B) B 2|x| 0
Z |
1 1
≥ x| dy
2|x| 0 y(log 1/y)2
1 1
= .
2|x| log(1/|x|)
Since f ∗ (x) is the supremum of such integrals over all balls containing
x, it is at least equal to the integral over B, so f ∗ (x) ≥ 2|x| log(1/|x|)
1
.
This function is not locally integrable, because if we integrate it in any
neighborhood around 0 we get
δ
Z δ  
1 1
dx = − log log = ∞.
−δ 2|x| log(1/x) x
0

Exercise 6: In one dimension there is a version of the basic inequality (1) for
the maximal function in the form of an identity. We defined the “one-sided”
maximal function
1 x+h
Z
∗
f+ (x) = sup |f (y)|dx.
h>0 h x
∗
If Eα+ = {x ∈ R : f+ (x) > α}, then
Z
1
m(Eα+ ) = |f (y)|dy.
α Eα+
38

Solution. First, we note that

1 x+h
Z
x ∈ Eα+ ⇔ ∃h > 0 s.t. |f (y)|dy > α
h x
Z x+h
⇔ |f (y)|dy > αh
x
Zx+h Z x
⇔ |f (y)|dy − |f (y)|dy − α(x + h) + αx > 0
0 0
Z x+h Z x
⇔ |f (y)|dy − α(x + h) > |f (y)|dy − αx.
0 0
Rx
Thus, if we define F (x) = 0 |f (y)|dy − αx, the set Eα+ is precisely the
set {x : ∃h > 0 s.t. F (x + h) > F (x)}. Note also that F is continuous
by the absolute continuity of the integral (we are assuming that f ∈ L1 ,
naturally). By the Rising Sun Lemma,
∞
[
Eα+ = (aj , bj )
j=1

where the intervals (aj , bj ) are disjoint and

F (aj ) = F (bj )
Z aj Z bj
⇒ |f (y)|dy − αaj = |f (y)|dy − αbj
0 0
Z bj
⇒ |f (y)|dy = α(bj − aj ).
aj

Then
Z ∞ Z
X bj ∞
X
|f (y)|dy = |f (y)|dy = α (bj − aj ) = αm(Eα+ )
+
Eα j=1 aj j=1

as desired. 

Exercise 8: Suppose A is a Lebesgue measurable set in R with m(A) > 0.

Does there exist a sequence {sn } such that the complement of ∪∞
n=1 (A+sn )
in R has measure zero?

Solution. Yes. Let x be any point of density of A, and let sn = qn − x

where {qn } is an enumeration of the rationals. Let E = ∪(A + sn ). To
show that m(E c ) = 0, it is sufficient to show that m(E c ∩ [n, n + 1]) = 0
for all n ∈ Z; since E is invariant under rational translations, it is sufficient
to show that E c ∩ [0, 1] has measure zero.
For m = 1, 2, . . . let Nm be an integer such that m(A ∩ B1/Nm (x)) ≥ (1 −
1
m )m(B1/Nm )(x). Such an Nm must exist because x is a point of density of
1
A. Then by the construction of E, m(A∩B1/Nm (q)) ≥ (1− m )m(B1/Nm )(q)
for any q ∈ Q. Now the open balls
 
j j
Um = B1/Nm , j = 1, 2, . . . , 2Nm
2Nm
 39

cover [0, 1], so

2N
[m
E c ∩ [0, 1] ⊂ E c ∩ Um
j


j=1
and
2N
X m

2N
X m
1 1 2 4
m (E c ∩ [0, 1]) ≤ m E c ∩ Um
j
≤ j
m(Um ) = (2Nm ) = .
j=1 j=1
m m N m m
4
Since m(E c ∩[0, 1]) ≤ m for all m, m(E c ∩[0, 1]) = 0. Hence m(E c ) = 0. 

Note: It is not sufficient to construct a set whose Lebesgue points are

dense in R. Consider an open dense set of measure ≤  (e.g. put an interval
of length 2k around the kth rational). Then every point is a Lebesgue point
since the set is open, yet its complement has positive measure.
Exercise 9: Let F be a closed subset in R, and δ(x) the distance function
from x to F , that is
δ(x) = inf{|x − y| : y ∈ F }.
Clearly, δ(x + y) ≤ |y| whenever x ∈ F . Prove the more refined estimate
δ(x + y) = o(|y|) for a.e. x ∈ F ,
that is, δ(x + y)/|y| → 0 for a.e. x ∈ F .

Solution. We note that δ is a function of bounded variation on any interval;

in fact, in general we have Vab (δ) ≤ |b − a| because |δ(y) − δ(z)| ≤ |y − z|
for all y, z ∈ R. Since δ has bounded variation, it is differentiable almost
everywhere; in particular, it is differentiable for a.e. x ∈ F . But δ is a
nonnegative function that is 0 for all x ∈ F , so it has a local minimum at
every point of F ; if it is differentiable at x ∈ F , its derivative is zero. By
the definition of derivative, this implies δ(x + y)/|y| → 0. 

Exercise 15: Suppose F is of bounded variation and continuous. Prove that

F = F1 − F2 , where both F1 and F2 are monotonic and continuous.

Solution. Every bounded-variation function is a difference of increasing

functions, so write F = G1 − G2 where G1 and G2 are increasing. As
shown in Lemmas 3.12-13, an increasing function is a continuous increasing
function plus a jump function. Hence G1 = F1 + J1 where F1 is continuous
and increasing, and J1 is a jump function; similarly, G2 = F2 + J2 . Then
F = (F1 − F2 ) + (J1 − J2 ). But J1 − J2 is a jump function, and jump
functions are continuous only if they’re constant. Since F is continuous,
this implies that J1 − J2 is constant; WLOG, J1 − J2 = 0. (Otherwise we
could redefine F10 = F1 + (J1 − J2 ) and F10 would also be continuous and
increasing.) Hence F = F1 − F2 . 

Exercise 17: Prove that if {K }>0 is a family of approximations to the

identity, then
sup|(f ∗ K )(x)| ≤ cf ∗ (x)
>0
for some constant c > 0 and all integrable f .
40

Exercise 18: Verify the agreement between the two definitions given for the
Cantor-Lebesgue function in Exercise 2, Chapter 1 and Section 3.1 of this
chapter.

Solution. This is such a lame problem. It’s so clear that they’re the same.
Probably the easiest way to see that is to think of the Cantor-Lebesgue
function as the following process:
• Given x, let y be the greatest member of the Cantor set such that
y ≤ x. (We know such a y exists because the Cantor set is closed.)
• Write the ternary expansion of y.
• Change all the 2’s to 1’s and re-interpret as a binary expansion. The
value obtained is F (x).
It’s pretty clear that both the definitions of the Cantor-Lebesgue function
given in the text do exactly this. 

Exercise 19: Show that if f : R → R is absolutely continuous, then

(a) f maps sets of measure zero to sets of measure zero.
(b) f maps measurable sets to measurable sets.

Solution.
(a) Suppose E ⊂ R hasPmeasure zero. Let P> 0. By absolute continuity,
∃δ > 0 such that |bj − aj | < δ ⇒ |f (bj ) − f (aj )| < . Since
m(E) = 0, there is an open set U ⊃ E with m(U ) < δ. Every open
subset of R is a countable disjoint union of open intervals, so
∞
[ ∞
X
U= (aj , bj ) with (bj − aj ) < δ.
j=1 j=1

For each j = 1, 2, . . . ,, let mj , Mj ∈ [aj , bj ] be values of x with

f (mj ) = min f (x) and f (Mj ) = max f (x).
x∈[a,b] x∈[a,b]

Such mj and Mj must exist because f is continuous and [aj , bj ] is

compact. Then
∞
[
f (U ) ⊂ [f (mj ), f (Mj )].
j=1

But |Mj − mj | ≤ |bj − aj | so

∞
X ∞
X
|Mj − mj | < δ ⇒ |f (Mj ) − f (mj )| < .
j=1 j=1

Hence f (E) is a subset of a set of measure less that . This is true for
all , so f (E) has measure zero.
(b) Let E ⊂ R be measurable. Then E = F ∪ N where F is Fσ and
N has measure zero. Since closed subsets of R are σ-compact, F is
σ-compact. But then f (F ) is also σ-compact since f is continuous.
Then f (E) = f (F ) ∪ f (N ) is a union of an Fσ set and a set of measure
zero. Hence f (E) is measurable.

 41

Exercise 20: This exercise deals with functions F that are absolutely con-
tinuous on [a, b] and are increasing. Let A = F (a) and B = F (b).
(a) There exists such an F that is in addition strictly increasing, but such
that F 0 (x) = 0 on a set of positive measure.
(b) The F in (a) can be chosen so that there is a measurable subset E ⊂
[A, B], m(E) = 0, so that F −1 (E) is not measurable.
(c) Prove, however, that for any increasing absolutely continuous F , and
E a measurable subset of [A, B], the set F −1 (E) ∩ {F 0 (x) > 0} is
measurable.
Solution.
(a) Let Z x
F (x) = δC (x)dx
a
where C ⊂ [a, b] is a Cantor set of positive measure and δC (x) is
the distance from x to C. Note that δC (x) ≥ 0 with equality iff
x ∈ C. Since δC is continuous, this integral is well-defined, even in the
Riemann sense. Moreover, F is absolutely continuous by the absolute
continuity of integration of L1 functions. As shown in problem 9, F 0 (x)
exists and equals zero a.e. in C, hence on a set of positive measure.
However, F is strictly increasing: Suppose a ≤ x < y ≤ b. Since C
contains no interval, some point, and therefore some interval, between
x and y belongs to C C . The integral of δC over this interval will be
positive, so F (y) > F (x).
(b) The same function from part (a) does the trick. Since F is increasing,
it maps disjoint open intervals to disjoint open intervals. Let U =
[a, b] \ C. Since U is open, we can write
∞
[
U= (aj , bj )
j=1

where the intervals (aj , bj ) are disjoint. Then

∞
[
F (U ) = (F (aj ), F (bj ))
j=1

and
∞
X
m(F (U )) = F ((bj ) − F (aj )).
j=1
But
Z b Z ∞
X
B − A = F (b) − F (a) = δ(x)dx = δ(x)dx = (F (bj ) − F (aj ))
a U j=1
R
since δ = 0 on C so C δ(x)dx = 0. Thus m(F (U )) = m(F ([a, b])), so
that m(F (C)) = 0. This implies that m(F (S)) = 0 for any subset S ⊂
C. But since C has positive measure, it has a non-measurable subset.
Then if E = F (S), m(E) = 0 so E is measurable, but F −1 (E) = S is
not measurable.
(c)

42

Exercise 22: Suppose that F and G are absolutely continuous on [a, b]. Show
that their product F G is also absolutely continuous. This has the following
consequences.
(a) Whenever F and G are absolutely continuous in [a, b],
Z b Z b
0
F (x)G(x)dx = − F (x)G0 (x)dx + [F (x)G(x)]ba .
a a

(b) Let F be absolutely continuous in [−π, π] with F (π) = F (−π). Show

that if Z π
1
an = F (x)e−inx dx,
2π −π
an einx , then
P
such that F (x) ∼
X
F 0 (x) ∼ inan einx .

(c) What happens if F (−π) 6= F (π)?

Proof. Since F and G are absolutely continuous, they are continuous and
therefore bounded on the compact interval [a, b]. Suppose |F |,P |G| ≤ M on
this interval. Now given  > 0, we can choose δ > 0 such that |bj − aj | <
 
P P
δ ⇒ |F (bj ) − F (aj ) < M and |G(bj ) − G(aj )| < 2M . Then
X
|F (bj )G(bj ) − F (aj )G(aj )|
X1
= |(F (bj − F (aj ))(G(bj ) + G(aj )) + (F (bj ) + F (aj ))(G(bj ) − G(aj ))|
2
1 X X 
≤ |F (bj ) − F (aj )||G(bj ) + G(aj )| + |F (bj ) + F (aj )||G(bj ) − G(aj )|
2
1 X X 
≤ (2M )|F (bj ) − F (aj )| + (2M )|G(bj ) − G(aj )|
2
1   
≤ 2M · + 2M · = .
2 2M 2M
This proves that F G is absolutely continuous on [a, b]. We now turn to the
consequences of this:
(a) Since F G is absolutely continuous, it’s differentiable almost every-
where. By elementary calculus, (F G)0 = F 0 G + F G0 at any point
where all three derivatives exist, which is almost everywhere. Inte-
grating both sides and subtracting F G0 yields
R

Z b Z b Z b
F 0 (x)G(x)dx = − F (x)G0 (x)dx + (F G)0 (x)dx.
a a a

Since F G is absolutely continuous, this implies

Z b Z b
0
F (x)G(x)dx = − F (x)G0 (x)dx + [F (x)G(x)]ba .
a a

(b) It would be nice if the problem would actually define this for us, but
I’m assuming that the ∼ here means “is represented by” as opposed to
any kind of statement about whether the function actually converges
 43

to its Fourier series or not. Then suppose bn are the Fourier coefficients
of F 0 , so by definition
Z π
1
bn = F 0 (x)e−inx dx.
2π −π
Using part (a), we have
Z π Z π
1 −inx 1
bn = − F (x)(−ine )dx+[F (x)e−inx ]ba = in F (x)e−inx dx = inan .
2π −π 2π −π

(c) Then all bets are off. As one example, consider F (x) = x which is
clearly absolutely continuous on [−π, π]. Then
Z π  π
1 xe−inx e−inx

1 −inx 2i
an = xe dx = + 2
= (−1)n
2π −π 2π −in n n
−π
Rπ
for n 6= 0, and a0 = −π xdx = 0. However, F 0 (x) = 1 which has
Fourier coefficients b0 = 1 and bn = 0 for n 6= 0.


Exercise 25: The following shows the necessity of allowing for general ex-
ceptional sets of measure zero in the differentiation Theorems 1.4, 3.4, and
3.11. Let E be any set of measure zero in Rd . Show that:
(a) There exists a non-negative integrable f in Rd , such that
Z
1
lim inf f (y)dy = ∞ for each x ∈ E.
m(B)→0 m(B) B
x∈B

(b) When d = 1 this may be restated as follows. There is an increasing

absolutely continuous function F such that
D+ F (x) = D− F (x) = ∞, for each x ∈ E.

Solution.
(a) Since E has measure zero, thereP exist open sets On with E ⊂ On for
∞
all n and m(On ) < 21n . Let f = n=1 χOn . Then f ∈ L1 since
Z ∞ Z ∞ ∞
X X X 1
f= χOn = m(On ) ≤ = 1.
Rd n=1 R
d
n=1 n=1
2n
Now let x ∈ E. Since On is open, there exist open balls Bn ⊂ On with
x ∈ Bn . Then for any ball B 3 x,
∞ ∞ ∞
m(Bn ∩ B)
Z Z
X X 1 X
f (y)dy = m(On ∩B) ≥ m(Bn ∩B) ⇒ f (y)dy ≥ .
B n=1 n=1
m(B) B n=1
m(B)

For any N , there exists δ > 0 such that m(B) < δ and x ∈ B implies
B ⊂ Bj for all j = 1, . . . , N . (This is true because B1 ∩ · · · ∩ Bn is
an open set containing x and hence contains an open ball around x.)
Then
∞
m(Bn ∩ B)
Z
1 X
f (y)dy ≥ ≥N
m(B) B n=1
m(B)
44

for sufficiently small B. This proves that

Z
1
lim inf f (y)dy = ∞.
m(B)→0 m(B) B
x∈B

(b) Let f be as in part (a), and

Z x
F (x) = f (y)dy.
0

Then F is absolutely continuous because it is the integral of an L1

function; it is increasing because f is nonnegative. Now
Z x+h
F (x + h) − F (x)
D+ F (x) = lim inf = lim inf f (y)dy
h→0
h>0
h h→0
h>0 x

and
x+h
F (x + h) − F (x)
Z
D− F (x) = lim inf = lim inf f (y)dy
h→0
h<0
h h→0
h<0 x

The conclusion in (a) implies that both of these are infinite, since
one can consider integrals over the balls [x, x + h) and (x − h, x].
(Technically, I suppose we should work with open balls, but one can
look at e.g. (x − δ, x + h) for sufficiently small δ.)


Exercise 30: A bounded function F is said to be of bounded variation on R if

F is of bounded variation on any finite sub-interval [a, b] and supa,b TF (a, b) <
∞. Prove
R that such an F enjoys the following two properties:
(a) RR |F (x + h) − F (x)|dx ≤ A|h|, for some constant A and all h ∈ R.
(b) | R F (x)φ0 (x)dx| ≤ A, where φ ranges over all C 1 functions of bounded
support with supx∈R |φ(x)| ≤ 1.
For the converse, and analogues in Rd , see Problem 6* below.

Solution.
(a) First, note that it is sufficient to treat the case where F is a bounded
increasing function. This is so because in general we can let F =
F1 − F2 where F1 and F2 are bounded increasing functions; if they
both satisfy the given condition, with constants A1 and A2 , then
|F (x + h)R − F (x)| ≤ |F1 (x + h) − F1 (x)| + |F2 (x + h) − F2 (x)| for
all x, so |F (x + h) − F (x)| ≤ A1 + A2 .
(In case someone asks why F must be a difference of bounded in-
creasing functions, we could re-do the proof that was used on finite
intervals, using the positive and negative variations of f . This avoids
the problem of trying to extend from the bounded case and worrying
about whether such an extension is unique.)
Suppose now that F is bounded and increasing. Since F is increas-
ing, |F (x + h) − F (x)| = F (x + h) − F (x) for h > 0. (By transla-
tion
R invariance, it is sufficient
R to treat theRcase of positive h, since
|F (x − h) − F (x)| = F (x) − F (x − h) = F (x + h) − F (x).) Then
 45

on any interval [a, a + h],

Z a+h
F (x+h)−F (x) ≤ F (a+2h)−F (a) ⇒ F (x+h)−F (x) ≤ h(F (a+2h)−F (a)).
a

In particular,
Z (n+1)h
≤ h(F ((n + 2)h) − F (nh))
nh

for any n ∈ Z. Then for any N ,

Z Nh N
X −1
F (x + h) − F (x) ≤ h F ((n + 2)h) − F (nh)
−N h n=−N
 
≤ h F ((N + 1)h) + F (N h) − F ((−N + 1)h) − F (−N h)
≤ 2h(F (+∞) − F (−∞))
since the sum telescopes. Since F (+∞) − F (−∞) is a finite constant,
this proves the result.
(b) By some algebraic legwork,



F (x + h)φ(x + h) − F (x)φ(x) = F (x) φ(x + h) − φ(x) − φ(x + h) F (x + h) − F (x) .
When we integrate both sides over R, the left-hand side integrates to
zero by translation invariance. Hence
Z Z



F (x) φ(x + h) − φ(x) dx = φ(x + h) F (x + h) − F (x) dx.
R R

By part (a), we have

Z Z



F (x) φ(x + h) − φ(x) dx = φ(x + h) F (x + h) − F (x) dx
R R
Z
≤ |φ(x + h)| F (x + h) − F (x) dx
ZR
≤ F (x + h) − F (x) dx
R
≤ A|h|.
Hence
φ(x + h) − φ(x)
Z
F (x) dx ≤ A.
R h
Now φ is supported on some compact set K, so φ0 is a continuous
function which is 0 outside K. Hence it has a maximum M . By the
Mean Value Theorem, any difference quotient of φ is at most M in
absolute value. Then if L is a bound for F on K, F (x) φx+h−φ(x)
h
is dominated by M LχK for all h. Hence we can use the Dominated
Convergence Theorem as h → 0 to obtain
Z
F (x)φ0 (x)dx ≤ A.
R


46

Exercise 31: Let F be the Cantor-Lebesgue function described in Section

3.1. Consider the curve that is the graph of F , that is, the curve given by
x(t) = t and y(t) = F (t) with 0 ≤ t ≤ 1. Prove that the length L(x̄) of the
segment 0 ≤ t ≤ x̄ of the curve is given by L(x̄) = x̄ + F (x̄). Hence the
total length of the curve is 2.

Solution. It is true for any increasing function with F (0) = 0 that L(x̄) ≤
x̄ + F (x̄), because for any partition 0 = t0 < t1 < · · · < tn = x̄,
n q
X n
X
(tj − tj−1 )2 + (F (tj ) − F (tj−1 ))2 ≤ (tj −tj−1 )+(F (tj )−F (tj−1 )) = x̄+F (x̄).
j=1 j=1

We wish to show that this upper bound is in fact the least upper bound
when F is the Cantor-Lebesgue function. Consider the iterates Fn (x) of
which this function is the limit. The interval [0, 1] can be divided into
2n+1 − 1 intervals on which Fn (x) alternately increases and stays constant;
suppose we label them I1 , C1 , I2 , C2 , . . . , C2n −1 , I2n . The intervals Cj have
varying lengths, since they correspond to intervals that are deleted from the
Cantor set at varying stages of the iteration; however, the Ij all have length
1
3n since they correspond to the intervals remaining in the nth iteration of
2n


the Cantor set. Hence the sum of the lengths of the Ij is 3n , while the
n

sum of the lengths of the Cj is 1 − 23n .
Now let x̄ ∈ [0, 1], and consider the partition Pn consisting of all points less
than or equal to x̄ which are an endpoint of one of the Cj or Ij . Thus we
have 0 = t0 < t1 < · · · < tm = x̄ where Fn is increasing on [t0 , t1 ], constant
on [t1 , t2 ], increasing on [t2 , t3 ], etc. Note also that F (tj ) = Fn (tj ) since all
the tj are endpoints of the Ck intervals, which remain fixed in all successive
iterations. Then
m q
X
(tj − tj−1 )2 + (F (tj ) − F (tj−1 ))2
j=1
Xm q m q
X
= (tj − tj−1 )2 + (F (tj ) − F (tj−1 ))2 + (tj − tj−1 )2 + (F (tj ) − F (tj−1 ))2
j=1 j=1
j odd j even
Xm m
X
≥ (F (tj ) − F (tj−1 )) + (tj − tj−1 )
j=1 j=1
j odd j even
X
=F (x̄) + |Ck ∩ [0, x̄]|
k
X
=F (x̄) + x̄ − |Ik ∩ [0, x̄]|
k
X
≥F (x̄) + x̄ − |Ik |
k
 n
2
=F (x̄) + x̄ − .
3
 47

Letting n → ∞, this approaches x̄ + F (x̄), which proves L(x̄) ≥ x̄ + F (x̄).

Since we already know L(x̄) ≤ x̄ + F (x̄), we have L(x̄) = x̄ + F (x̄) as
desired. 
Exercise 32: Let f : R → R. Prove that f satisfies the Lipschitz condition
|f (x) − f (y)| ≤ M |x − y|
for some M and all x, y ∈ R, if and only if f satisfies the following two
properties:
(i) f is absolutely continuous.
(ii) |f 0 (x)| ≤ M for a.e. x.

Solution. Suppose P
P f is Lipschitz. Then for any  > 0, if we let δ = M , then
|bj − aj | < δ ⇒ |f (bj ) − f (aj )| < . Hence f is absolutely continuous.
This implies f is differentiable a.e.; if x is a point for which f 0 (x) exists, the
Lipschitz condition implies | f (x+h)−f h
(x)
| ≤ M for all h. Taking the limit
as h → 0 implies |f 0 (x)| ≤ M .
Conversely, suppose f is absolutely continuous and has bounded deriva-
tive a.e. Since absolutely continuous functions are the integrals of their
derivatives,
Z y Z max(x,y) Z max(x,y)
0 0
|f (y) − f (x)| = f (t)dt ≤ |f (t)|dt ≤ M dt = M |x − y|
x min(x,y) min(x,y)

so f is Lipschitz. 

Chapter 3.6, Page 152

Problem 4: A real-valued function φ defined on an interval (a, b) is convex
if the region lying above its graph {(x, y) ∈ R2 : y > φ(x), a < x < b} is a
convex set. Equivalently, φ is convex if
φ(θx1 + (1 − θ)x2 ) ≤ θφ(x1 ) + (1 − θ)φ(x2 )
for every x1 , x2 ∈ (a, b) and 0 ≤ θ ≤ 1. One can also observe as a conse-
quence that we have the following inequality of the slopes:
φ(x + h) − φ(x) φ(y) − φ(x) φ(y) − φ(y − h)
(1) ≤ ≤ ,
h y−x h
whenever x < y, h > 0, and x + h < y. The following can then be proved.
(a) φ is continuous on (a, b).
(b) φ satisfies a Lipschitz condition of order 1 in any proper closed sub-
interval [a0 , b0 ] of (a, b). Hence φ is absolutely continuous in each sub-
interval.
(c) φ0 exists at all but an at most denumerable number of points, and
φ0 = D+ φ is an increasing function with
Z y
φ(y) − φ(x) = φ0 (t)dt.
x
(d) RConversely, if ψ is any increasing function on (a, b), then φ(x) =
x
c
ψ(t)dt is a convex function in (a, b) for c ∈ (a, b).
Solution.
48

(a) Suppose to the contrary that φ is discontinuous at some x ∈ (a, b).

This means there exists  > 0 and an infinite sequence xn → x with
xn ∈ (a, b) and |φ(xn ) − φ(x)| >  for all n. Since the sequence xn
is infinite, it must have infinitely many points in one of the following
categories:
(i) φ(xn ) > φ(x) + 
(ii) φ(xn ) < φ(x) − 
We will treat each case separately and obtain a contradiction.
(i) Assume WLOG that the entire sequence xn is in this category
(otherwise, take a subsequence that is). We may also assume xn
converges to x monotonically, since otherwise we can again take
a subsequence. Let L(θ) = θφ(x) + (1 − θ)φ(x1 ) for 0 ≤ θ ≤ 1.
This is a continuous function of θ, so ∃δ > 0 such that |θ| < δ ⇒
n −x1
L(θ) < φ(x) + . Now let θn = xx−x 1
. Then 0 ≤ θn ≤ 1 since
xn → x monotonically. Note also that xn = θn x + (1 − θn )x1 .
Now θn → 0 as n → ∞, so θn < δ for n sufficiently large. But
this implies
φ(xn ) = φ(θn x + (1 − θn )x1 ) ≤ L(θn ) < φ(x) + ,
for sufficiently large n, a contradiction.
(ii) Again, we assume WLOG that xn → x monotonically. Let y ∈
(a, b) such that x is between y and xn for all n. Then θn = xxnn −x
−y
has the properties that 0 ≤ θn ≤ 1 and θn → 1 as n → ∞. Now
x = θn xn + (1 − θn )y, so
φ(x) ≤ θn φ(xn ) + (1 − θn )φ(y)
for all n. But θn → 1, and since φ(xn ) < φ(x) − , this implies
φ(x) < φ(x) −  for sufficiently large n, a contradiction.
Hence φ is continuous.
(b) First, I prove an inequality of slopes that I like better than the one
given. I claim that for s < t < u with s, t, u ∈ (a, b),
φ(t) − φ(s) φ(u) − φ(s) φ(u) − φ(t)
(2) ≤ ≤ .
t−s u−s u−t
This follows straightforwardly from the convexity condition:
u−t t−s
t= s+ u
u−s u−s
u−t t−s
⇒ φ(t) ≤ φ(s) + φ(u)
u−s u−s
(3) ⇒ (u − s)φ(t) ≤ (u − t)φ(s) + (t − s)φ(u)
⇒ (u − s)φ(t) − (u − s)φ(s) ≤ (s − t)φ(s) + (t − s)φ(u)
φ(t) − φ(s) φ(u) − φ(s)
⇒ ≤ .
t−s u−s
Taking a different route from inequality (3) leads to
(u − t)φ(u) − (u − t)φ(s) ≤ (u − s)φ(u) − (u − s)φ(t)
φ(u) − φ(s) φ(u) − φ(t)
⇒ ≤
u−s u−t
 49

as desired. Given this inequality of slopes, we can easily prove that φ

is Lipschitz on [a0 , b0 ]. Choose h > 0 such that [a0 − h, b0 + h] ⊂ (a, b).
Then for x, y ∈ [a0 , b0 ], suppose WLOG that x < y; since a0 − h < a0 ≤
x < y ≤ b0 < b0 + h, the slope inequality yields

φ(a0 − h) − φ(a0 ) φ(y) − φ(a0 − h) φ(y) − φ(x) φ(b0 + h) − φ(x) φ(b0 + h) − φ(b0 )
≤ ≤ ≤ ≤ .
h y − a0 + h y−x b0 + h − x h

The leftmost and rightmost terms above are constants, which we may
call m and M ; we thus have m|y − x| ≤ |φ(y) − φ(x)| ≤ M |y − x|,
whence φ is Lipschitz on [a0 , b0 ].
(c) Since φ is Lipschitz on any closed subinterval [x, y] ⊂ (a, b), φ is abso-
Rlutely continuous on [x, y] by Exercise 32 above; hence φ(y) − φ(x) =
y 0
x
φ (t)dt.
Now inequality (2) implies that φ(x+h)−φ(x) h is an increasing function
of h at any x ∈ (a, b). This implies that D+ = D+ and D− = D− ,
that |D+ |, |D− | < ∞, and that D+ ≥ D− . The inequality (1) tells
us that x < y ⇒ D+ φ(x) ≤ D− φ(y). This in turn implies that D+
and D− are increasing. To show that D+ = D− except at countably
many points, let {xα } be those points in (a, b) for which this is not
true, and define jα > 0 by jα = D+ (φ)(xα ) − D− (φ)(xα ). Then on
any subinterval [a0 , b0 ], if {x1 , . . . , xn } ⊂ [a0 , b0 ], we have

n+1
X 
D+ (φ)(xk ) − D− (φ)(xk )
k=1
n+1
X  n+1
X 
≤ D+ (φ)(xk ) − D− (φ)(xk ) + D− (φ)(xk ) − D+ (φ)(xk−1 )
k=1 k=1
=D+ (φ)(b0 ) − D− (φ)(a0 ),

where we use the convention x0 = a0 and xn+1 = b0 . This implies that

X  
D+ (φ)(xk ) − D− (φ)(xk )
x∈[a0 ,b0 ]

is finite, because all finite sub-sums are bounded by the finite constant
D+ (φ)(b0 )−D− (φ)(a0 ). So this sum can containly only countably many
nonzero terms, which means only countably many points in [a0 , b0 ] can
have D− (φ)(x) 6= D+ (φ)(x). Since (a, b) is a countable union of closed
subintervals (e.g. ∩[a + n1 , b − n1 ]), it can contain only countably many
points for which D+ 6= D− . Everywhere else, the derivative exists.
50

(d) Because ψ is increasing,

Z θx1 +(1−θ)x2


φ θx1 + (1 − θ)x2 = ψ(t)dt
Zc x2 Z x2
= ψ(t)dt − ψ(t)dt
c θx1 +(1−θ)x2
Z x2 Z x2 Z x2
= ψ(t)dt − θ ψ(t)dt − (1 − θ) ψ(t)dt
c θx1 +(1−θ)x2 θx1 +(1−θ)x2
Z x2 Z x2
≤ ψ(t)dt − θ ψ(t)dt − (1 − θ)θ(x2 − x1 )ψ(θx1 + (1 − θ)x2 )
c θx1 +(1−θ)x2
Z x2 Z x2 Z θx1 +(1−θ)x2
≤ ψ(t)dt − θ ψ(t)dt − θ ψ(t)dt
c θx1 +(1−θ)x2 x1
Z x2 Z x2
= ψ(t)dt − θ
ψ(t)dt
c x1
Z x1 Z x2
=θ ψ(t)dt + (1 − θ) ψ(t)dt
c c
= θφ(x1 ) + (1 − θ)φ(x2 ).
So φ is convex.


Chapter 4.7, Page 193

Exercise 4: Prove from the definition that `2 (Z) is complete and separable.

Solution. The proof that `2 is complete is exactly the same as the proof
(m)
that L2 is complete, from pp. 159-160 of the textbook. Let {aj }∞ m=1 be
2
a Cauchy sequence in ` (Z). For each k ≥ 1 we can choose nk such that
m, n ≥ nk ⇒ ka(m) − a(n) k < 21k and nk < nk+1 . Then the subsequence
a(nk ) has the property that ka(nk+1 ) − a(nk ) k ≤ 21k . Define sequences a =
{aj } and b = {bj } by
∞  
(n1 ) (nk+1 ) (nk )
X
aj = aj + aj − aj
k=1

and
∞
(n1 ) (nk+1 ) (nk )
X
bj = |aj |+ aj − aj
k=1
and the partial sums
K  
(a,K) (n1 ) (nk+1 ) (nk )
X
Sj = aj + aj − aj
k=1

and
K
(b,K) (n1 ) (nk+1 ) (nk )
X
Sj = |aj |+ aj − aj .
k=1
 51

Then
K
X 1
kS (b,K) k ≤ ka(n1 ) k +
2k
k=1
by the triangle inequality; letting K → ∞, kbk converges by the monotone
convergence theorem (for sums, but hey, sums are just integrals with dis-
crete measures), so kak converges since it converges absolutely. Of course,
this implies that bj and hence aj converges for each j; since the partial
(n )
sums are S (a,K) = a(nk+1 ) by construction, aj k+1 → aj for all j. Now
given  > 0, choose N such that ka(n) − a(m) k < 2 for n, m > N , and let
nK > N such that ka(nK ) − ak < 2 . Then
m > N ⇒ ka(m) − ak ≤ ka(m) − a(nK ) k + ka(nK ) − ak < .
Hence a(m) → a.
To prove that `2 is separable, consider the subset D consisting of all rational
sequences which are 0 except at finitely many values. This is countable
because
D = {rational sequences of finite length}
∞
[
= {sequences {an } with an ∈ Q and an = 0 for |n| > N }
N =1
[∞
= Q2N +1
N =1

is a countable union of countable sets. Now let {bn } ∈ `2 be any square

summable sequence. Given  > 0, there exists N such that
X 2
|bn |2 <
2
|n|>N

since the infinite sum converges. Then for each j = −N, . . . , N , we can
2
choose a rational number qj with |qj − bj |2 < 22+N +j . If we also define

qj = 0 for |j| > N , then q ∈ D and

∞ N 2N
X X X 2 X 2
kq − bk2 = |qj − bj |2 = |bj − 0|2 + |qj − bj |2 < + 2+s
< 2 .
j=−∞
2 s=0
2
|j|>N j=−N

This shows that D is dense. 

Exercise 5: Establish the following relations between L2 (Rd ) and L1 (Rd ):
(a) Neither the inclusion L2 (Rd ) ⊂ L1 (Rd ) nor the inclusion L1 (Rd ) ⊂
L2 (Rd ) is valid.
(b) Note, however, that if f is supported on a set E of finite measure and
if f ∈ L2 (Rd ), applying the Cauchy-Schwarz inequality to f χE gives
f ∈ L1 (Rd ), and
kf k1 ≤ m(E)1/2 kf k2 .
(c) If f is bounded (|f (x)| ≤ M ), and f ∈ L1 (Rd ), then f ∈ L2 (Rd ) with
1/2
kf k2 ≤ M 1/2 kf k1 .
52

Solution.
(a) Let f (x) = χ|x|≥1 |x|1d/2 and g(x) = χ|x|≤1 |x|1 d . Then Exercise 10 of
Chapter 2 shows that f and g 2 are integrable, but f 2 and g are not.
(b) Applying the Cauchy-Schwarz inequality to the inner product of f χE
and χE ,
Z Z 1/2 Z 1/2
kf k1 = kf χE k1 = |f χE | ≤ |f | χE = m(E)1/2 kf k2 .

(c) Since |f | ≤ M ,
Z Z
1/2
|f |2 ≤ M |f | ⇒ kf k22 = |f |2 ≤ M |f | = M kf k1 ⇒ kf k2 ≤ M 1/2 kf k1 .

Exercise 6: Prove that the following are dense subspaces of L2 (Rd ):

(a) The simple functions.
(b) The continuous functions of compact support.

Solution.
(a) It is sufficient to treat the case of nonnegative f , since every complex
L2 function is a linear combination of nonnegative L2 functions. We
know there exists a sequence of simple functions sn % f with 0 ≤ sn ≤
p p
Rf . Then |fp − sn | ≤ |f | so by the Dominated Convergence Theorem,
|f − sn | → 0. Hence sn → f in Lp . Therefore the simple functions
are dense.
(b) Let s ∈ Lp (Rd ) be a simple function. It is sufficient to find g ∈
CC (Rd ) with kg − skp < . If s = 0, s ∈ CC (Rd ) and we’re done.
Otherwise, since s is simple, 0 < ksk∞ < ∞; since it is in Lp , it
must be supported on a set E of finite measure. Now Lusin’s theorem
p
d 
enables us to construct g ∈ CC (R ) with m({g 6= s}) < 2ksk∞ and
sup |g| ≤ sup |s| < ∞. (To do this, construct g from Lusin’s theorem;
then if g ≥ ksk∞ on the closed set F , change g to ksk∞ on F . See
Rudin, Real and Complex Analysis pp. 55-56.) Then |g − s| ≤ 2ksk∞
 p

and is nonzero on a set of measure 2ksk∞ , so
Z  p
p p 
|g − s| ≤ 2 ksk∞ = p ⇒ kg − skp < .
2ksk∞


Exercise 7: Suppose {φk }∞ 2 d

k=1 is an orthonormal basis for L (R ). Prove that
the collection {φk,j }1≤k,j<∞ with φk,j (x, y) = φk (x)φj (y) is an orthonormal
basis of L2 (Rd × Rd ).

Solution. First, note that φk,j is indeed in L2 (Rd × Rd , since

Z Z Z  Z 
2 2 2 2 2
|φ(k, j)| = |φk (x)| |φj (y)| dxdy = |φk (x)| dx |φj (y)| dy = 1
R2d R2d Rd Rd
 53

by Fubini’s Theorem. Also,

Z
hφj,k φ`,m i = φj,k φ`,m
2d
ZR
= φj (x)φk (y)φ` (x)φm (y)dxdy
R2d
Z  Z 
= φj (x)φ` (x)dx φk (y)φm (y)dy
Rd Rd
` m
= δj δk
so {φj,k } is an orthonormal set. To show that linear combinations of {φj,k }
are dense in L2 (Rd ), one approach would be to blow this problem out
of the water with the Stone-Weierstrass theorem. We know CC (Rd ) is
dense in L2 (Rd ), and the Stone-Weierstrass theorem tells us that linear
combinations
Pm of “separable” continuous functions (i.e. functions of the form
f
i=1 i (x)g i (y)) are dense in CC (Rd ). It is easy to verify that functions of
this form can be approximated by {φj,k }, so we’re done.
Alternatively, we can follow the approach given in the hint. Let f ∈ L2 (R2d )
and suppose that hf, φj,k i = 0 for all j and k. By Fubini’s Theorem,
Z
0= f (x, y)φj,k (x, y)
R2d
Z Z 
= f (x, y)φj,k (x, y)dy dx
Rd Rd
Z Z 
= f (x, y)φk (y)dy φj (x)dx
Rd Rd

Hence, if we define
Z
fk (x) = f (x, y)φk (y)dy,
Rd
we see that Z
fk (x)φj (x) = 0
Rd
for all j. Because {φj } is an orthonormal basis, this implies that fk (x) = 0
for all k. Because {φk } is an orthonormal basis, this in turn implies that
f (x, y) = 0. Since f ⊥ φj,k ⇒ f = 0, {φj,k } is an orthonormal basis. 

Exercise 9: Let H1 = L2 ([−π, π]) be the Hilbert space of functions F (eiθ )

1
Rπ
on the unit circle with inner product (F, G) = 2π −π
F (eiθ )G(eiθ )dθ. Let
2
H2 be the space L (R). Using the mapping
i−x
x 7→
i+x
of R to the unit circle, show that:
(a) The correspondence U : F → f , with
 
1 i−x
f (x) = 1/2 F
π (i + x) i+x
gives a unitary mapping of H1 to H2 .
54

(b) As a result,
  n ∞
1 i−x 1
π 1/2 i+x i+x n=−∞
2
is an orthonormal basis of L (R).
Solution.
(a) If we define θ = 2 tan−1 (x), then x = tan θ2 , i−x

iθ 2
i+x = e , 1 + x =
sec2 θ2 , and dx = 12 sec2 θ2 dθ. (Brings back memories of high school

calculus, don’t it?) Then

 2
i−x
Z Z
2 1
|f (x)| dx = 2
F dx
R R π|i + x| i+x
 2
i−x
Z
1 1
= 2
F dx
R πx +1 i+x
Z π  
1 1 iθ 2 1 2 θ
= 2
|F (e )| sec dθ
−π π sec (θ/2) 2 2
Z π
1
= |F (eiθ )|2 dθ
2π −π
so kf kH2 = kF kH1 . So U is unitary.
(b) By the Riesz-Fischer theorem, {einθ } is an orthonormal basis for L2 (T ).
Because U is unitary,
  n 
inθ 1 i−x 1
{U (e )} = √
π i+x i+x
is an orthonormal basis for L2 (R).

Exercise 10: Let S denote a subspace of a Hilbert space H. Prove that
(S ⊥ )⊥ is the smallest closed subspace of H that contains S.
Solution. Let \
S̄ = V.
V ⊂H subspace
V closed

Then S̄ is a closed subspace, because the intersection of closed sets is closed

and the intersection of subspaces is a subspace. It is obviously the smallest
closed subspace containing S. We want to show that S̄ = (S ⊥ )⊥ . Clearly
S̄ ⊂ (S ⊥ )⊥ since the latter is a closed subspace containing S. To establish
the reverse inclusion, we first show that (S̄)⊥ = S ⊥ . Clearly (S̄)⊥ ⊂ S ⊥
since S ⊂ S̄. To show the opposite inclusion, let x ∈ S ⊥ . Then for any
w ∈ S̄, there is a sequence wn ⊂ S with wn → w. Because inner products
are continuous, 0 = hwn , xi → hw, xi so x ⊥ w for any w ∈ S̄. This proves
S ⊥ ⊂ (S̄)⊥ . Since they’re equal, we can use Proposition 4.2 to write
H = S̄ ⊕ S ⊥ .
Now let x ∈ (S ⊥ )⊥ . Then we can write x = v +w where v ∈ S̄ and w ∈ S ⊥ .
Then
hx, wi = hv, wi + hw, wi = hw, wi.
 55

But hx, wi = 0 because w ∈ S ⊥ and x ∈ (S ⊥ )⊥ . Hence w = 0 and

x ∈ S̄. 
Exercise 11: Let P be the orthogonal projection associated with a closed
subspace S in a Hilbert space H, that is,
P (f ) = f if f ∈ S and P (f ) = 0 if f ∈ S ⊥ .
(a) Show that P 2 = P and P ∗ = P .
(b) Conversely, if P is any bounded operator satisfying P 2 = P and P ∗ =
P , prove that P is the orthogonal projection for some closed subspace
of H.
(c) Using P , prove that if S is a closed subspace of a separable Hilbert
space, then S is also a separable Hilbert space.
Solution.
(a) Let x ∈ H and write x = xS + xS ⊥ . Then
P 2 (x) = P (P (x)) = P (xS ) = xS = P (x)
so P 2 = P . Moreover, if y = yS + yS ⊥ is any other vector in H, then
hP x, yi = hxS , yS + yS ⊥ i = hxS , yS i = hxS + xS ⊥ , yS i = hx, P yi
so P = P ∗ .
(b) Let S = im(P ) which is a subspace of H. To show S is closed, suppose
xn ∈ S and xn → x. Then because P is bounded, it’s continuous,
so P xn → P x. But P xn = xn , so xn → P x which implies P x = x.
Hence x ∈ S, so S is closed. Also, if w ∈ S ⊥ , then for all v ∈ S,
hx, P wi = hP x, wi = hx, wi = 0
so P w ∈ S ⊥ . But P w ∈ S, so P w = 0. Now using Proposition 4.2, if
y is any vector in H, then
y = yS + yS ⊥ .
Then by linearity, P (y) = P (yS ) + P (yS ⊥ ) = yS + 0 = yS . So P does
the same thing to y as orthogonal projection onto S, for any y ∈ H.
(c) Let H be a separable Hilbert space, {φn } a countable dense set, and S
a closed subspace. I claim that {PS φn } is dense in S. For any x ∈ S,
we can find a sequence φk → x. Then PS (φk − x) = PS φk − x since
P x = x. Since projections do not increase length,
kPS φk − xk ≤ kφk − xk ⇒ (PS φk ) → x.

d
Exercise 12: Let E be a measurable subset of R , and suppose S is the
subspace of L2 (Rd ) of functions that vanish for a.e. x ∈
/ E. Show that the
orthogonal projection P on S is given by P (f ) = χE · f , where χE is the
characteristic function of E.
Solution. Define a linear operator T : L2 (Rd ) → L2 (Rd ) by T (f ) = χE f .
Then T 2 (f ) = χ2E f = χE f = T (f ). Moreover,
Z Z
hT f, gi = χE f ḡ = f χE g = hf, T gi
Rd Rd
56

so T ∗ = T . We also note that T is bounded since |T f (x)| ≤ |f (x)| for all

x, so kT f k ≤ kf k. By problem 11c, T is a projection onto its image. But
im(T ) is precisely those functions which are 0 a.e. on E c . Hence T is the
desired projection. 
Exercise 13: Suppose P1 and P2 are a pair of orthogonal projections on S1
and S2 , respectively. Then P1 P2 is an orthogonal projection if and only if
P1 and P2 commute, that is, P1 P2 = P2 P1 . In this case, P1 P2 projects onto
S1 ∩ S2 .
Solution. Suppose P1 P2 is an orthogonal projection. Then
P2 P1 = P2∗ P1∗ = (P1 P2 )∗ = P1 P2
so they commute. On the other hand, suppose they commute; then
(P1 P2 )2 = P12 P22 = P1 P2
and
(P1 P2 )∗ = P2∗ P1∗ = P2 P1 = P1 P2
so P1 P2 is an orthogonal projection. The image of P1 P2 is a subspace of
S1 because P1 P2 v = P1 (P2 v) ∈ S1 for any v; similarly, it’s a subspace of S2
because P1 P2 v = P2 (P1 v) ∈ S2 . Hence the image of P1 P2 is a subspace of
S1 ∩ S2 . But every vector in S1 ∩ S2 is fixed by both P1 and P2 , and hence
by P1 P2 . Thus, the image of P1 P2 is precisely S1 ∩ S2 . 
Exercise 14: Suppose H and H0 are two completions of a pre-Hilbert space
H0 . Show that there is a unitary mapping from H to H0 that is the identity
on H0 .
Solution. Define U : H → H0 as follows: Given x ∈ H, choose a sequence
fn → x with fn ∈ H0 . Define U (x) = lim fn in H0 . (This limit exists
because fn is Cauchy.) To show this is well-defined, suppose gn ∈ H0 is
another sequence converging to x. Then fn −gn → 0 in both spaces (because
fn − gn ∈ H0 for all n), so they have the same limit in H0 as well. This
shows that U is well-defined. Clearly it is the identity on H0 . To show that
it’s unitary, we need only use the continuity of the norm: Suppose fn → x
in H and fn → U (x) in H0 . Then kxk = lim kfn k = kU (x)k. Hence U is
unitary. 
Exercise 15: Let T be any linear transformation from H1 to H2 . If we
suppose that H1 is finite-dimensional, then T is automatically bounded.
Solution. Let {e1 , . . . , en } be an orthonormal basis for H1 . (This is a basis
in the usual linear algebra sense, i.e. every vector is a finite linear combi-
nation of the basis vectors.) PnLet mi = kTP (ei )k. Then if x ∈ H1 is a unit
n
vector, we can write x = i=1 ci ei with i=1 |ci |2 = 1. By the Triangle
Inequality,
n
! n n
X X X
kT (x)k = T ci ei = ci T (ei ) ≤ |ci |kT (ei )k ≤ nM
i=1 i=1 i=1
√
where MP= max(m√ i ). (In fact, this bound can be improved to M n,
because |ci | ≤ n by the Cauchy-Schwarz inequality.) 
 57

Exercise 18: Let H denote a Hilbert space, and L(H) the vector space of all
bounded linear operators on H. Given T ∈ L(H), we define the operator
norm
kT k = inf{B : kT vk ≤ Bkvk, for all v ∈ H}.

(a) Show that kT1 + T2 k ≤ kT1 k + kT2 k whenever T1 , T2 ∈ L(H).

(b) Prove that
d(T1 , T2 ) = kT1 − T2 k

defines a metric on L(H).

(c) Show that L(H) is complete in the metric d.

Solution.
(a) This is easier if we use another expression for kT k, such as

kT xk
kT k = sup .
xinH,x6=0 kxk

If we call this supremum S, then we have kT xk ≤ Skxk for all x;

moreover, for any α < S, there exists x with kT xk/kxk > α ⇒ kT xk >
αkxk. Hence S is the infimum (in fact, the minimum) of such bounds,
so this definition really is equivalent. Then

kT1 xk kT2 xk kT1 xk kT2 xk

kT1 + T2 k = sup + ≤ sup + sup = kT1 k + kT2 k.
kxk kxk kxk kxk

(b) Since kT xk = k − T xk for all x, kT k = k − T k. Then

d(T1 , T2 ) = kT1 − T2 k = kT2 − T1 k = d(T2 , T1 ).

Clearly d(T1 , T1 ) = 0; conversely, if T1 6= T2 then there exists some x

with k(T1 − T2 )xk > 0, so kT1 − T2 k > 0. Finally, using part (a),

d(T1 , T3 ) = kT1 −T3 k = k(T1 −T2 )+(T2 −T3 )k ≤ kT1 −T2 k+kT2 −T3 k = d(T1 , T2 )+d(T2 , T3 ).

Hence d is a metric.
(c) Let {Tn } be a Cauchy sequence in L(H). Define T (x) = lim Tn (x) for
all x ∈ H. This limit exists because k(Tm − Tn )xk ≤ kTm − Tn kkxk
so {Tn x} is a Cauchy sequence in H. T is linear by the linearity of
limits. Finally, since Tn x → T x, the continuity of the norm implies
kTn xk → kT xk for all x. Hence kT k = lim kTn k which is finite because
|(kTm k − kTn k)| ≤ kTm − Tn k by the triangle inequality, so kTn k is a
Cauchy sequence of real numbers. So T is bounded.


Exercise 19: If T is a bounded linear operator on a Hilbert space, prove that

kT T ∗ k = kT ∗ T k = kT k2 = kT ∗ k2 .
58

Solution. We already know kT k2 = kT ∗ k2 from Proposition 5.4. Now

kT ∗ T k = sup |hT ∗ T f, gi|
kf k=kgk=1

= sup |hT f, T gi|

kf k=kgk=1

≤ sup kT f kkT gk
kf k=kgk=1

= sup kT f k sup kT gk
kf k=1 kgk=1
2
= kT k .
To show that equality is achieved, choose a sequence fn with kfn k = 1 and
kT fn k → kT k. Then hT fn , T fn i → kT k2 , so
kT ∗ T k = sup hT f, T gi ≥ sup hT f, T f i ≥ kT k2 .
kf k=kgk=1 kf k=1

Hence kT ∗ T k = kT 2 k. Finally, replacing T with T ∗ yields kT T ∗ k =

kT ∗ k2 = kT k2 and we are done. 
Exercise 20: Suppose H is an infinite-dimensional Hilbert space. We have
seen an example of a sequence {fn } in H with kfn k = 1 for all n, but
for which no subsequence of {fn } converges in H. However, show that for
any sequence {fn } in H with kfn k = 1 for all n, there exist f ∈ H and a
subsequence {fnk } such that for all g ∈ H, one has
lim (fnk , g) = (f, g).
n→∞

One says that {fnk } converges weakly to f .

Solution. The proof is similar to the Arzela-Ascoli theorem, and as with
that proof, the main hang-up is notation. Let {en } be an orthonormal
basis for H. (We’re assuming here that H is separable, of course, which the
book includes in its definition of Hilbert space.) Then we can define f by
defining its inner product with each en . Now hfn , e1 i is a sequence of real
numbers in [−1, 1] (by the Cauchy-Schwarz inequality, since these are both
unit vectors). Since [−1, 1] is compact, there is a subsequence fnj which
converges to some limit `1 . Then hfnj , e2 i is a sequence of real numbers in
[−1, 1], so some subsequence fnjr converges to a limit `2 . Continuing in this
fashion, we obtain sequences S0 , S1 , S2 , . . . where Sm+1 is a subsequence
(m) n (m)
of Sm , S0 = {fn }, and hvn , ej i → `j for j = 1, . . . , m, where vn is the
(k)
nth term of Sm . Define a sequence S whose kth term is vk . Then for any
k
Sm , the tail of S is a subsequence of the tail of Sm . Hence hvk , ej i → `j for
all j. Define f by hf, ej i = `j . Relabeling, vk = fnk is a subsequence with
hfnk , ej i → hf, ej i for all basis elements ej , and thus hfnk , gi → hf, gi for
all g ∈ H. 
Exercise 17: Fatou’s theorem can be generalized by allowing a point to ap-
proach the boundary in larger regions, as follows.
For each 0 < s < 1 and point z on the unit circle, consider the region
Γs (z) defined as the smallest closed convex set that contains z and the
closed disc Ds (0). In other words, Γs (z) consists of all lines joining z with
 59

pointsn in Ds (0). Near the point z, the region Γs (z) looks like a triangle.
See Figure 2.
We say that a function F defined in the open unit disc has a non-
tangential limit at a point z on the circle, if for every 0 < s < 1, the
limit
F (w)
w→z
w∈Γs (z)

exists.
Prove that if F is holomorphic and bounded on the open unit disc, then
F has a non-tangential limit for almost every point on the unit circle.
P∞ n
Solution. Since F is holomorphic, we have F (z) = n=0 aP n z for |z| < 1.
As shown on page 174 in the proof of Fatou’s theorem, |an |2 < ∞ so
2 iθ
there is an L (T ) function F (e ) whose Fourier coefficients are an . Note
also that F (eiθ ) is bounded (almost everywhere) since, by Fatou’s theorem,
it is the a.e. radial limit of F (z), so |F (z)| ≤ M ⇒ |F (eiθ )| ≤ M .
We next prove a lemma about the Poisson kernel.
Lemma 2. For each s ∈ (0, 1) there exists a constant ks such that
Pr (θ − φ) ≤ ks Pr (−φ)
for all (r, θ) such that reiθ ∈ Γs .
Proof. By elementary arithmetic,
1 − r2
Pr (θ − φ) =
|eiφ − reiθ |2
and
1 − r2
Pr (−φ) = .
|eiφ − r|2
(This alternate formula can be found in any complex analysis book.) Our
task is thus reduced to proving
|eiφ − r| ≤ ks |eiφ − reiθ |.
By the triangle inequality,
 
|θ|
|eiφ − r| ≤ |eiφ − reiθ | + r|eiθ − 1| = |eiφ − reiθ | + 2r sin .
2
Thus, our task is reduced to proving that
2r sin θ2

|eiφ − reiθ |
is bounded on Γs . But |eiφ − reiθ | ≥ 1 − r by the triangle inequality, so it
is sufficient to prove that
2r sin θ2

1−r
is bounded on Γs . Now for each r, the maximum value of |θ| (which will
maximize
this quotient) is, as indicated in the diagram
√ below,
√ one for which
2r sin θ2 occurs in a triangle with 1 − r and 1 − s2 − r2 − s2 . Thus,
60

it is at most equal to the sum of them (actually quite a bit less). So we

finally need only to prove that
√ √
1 − s2 − r2 − s2
1−r
is bounded as r → 1. But this follows from the fact that it has negative
derivative, since
d p p  r d
1 − s2 − r2 − s2 = − √ < −1 = (1 − r).
dr 2
r −s 2 dr
This completes the proof of the lemma. (I know, there are probably much
shorter proofs, but it’s late at night...) 
 61

Having established this lemma, the rest of the problem becomes trivial.
By the Poisson integral formula,
Z π
1
|F (reiθ )| = Pr (θ − φ)F (eiφ dφ
2π −π
Z π
1
≤ Pr (θ − φ)|F (eiφ |dφ
2π −π
Z π
1
≤ ks Pr (−φ)|F (eiφ |dφ
2π −π
and by Math 245A (specifically, the fact that Pr is an approximate identity)
this last integral tends to zero. (Recall that we assumed F (1) = 0.) 

Exercise 21: There are several senses in which a sequence of bounded oper-
ators {Tn } can converge to a bounded operator T (in a Hilbert space H).
First, there is convergence in the norm, that is, kTn − T k → 0 as n → ∞.
Next, there is a weaker convergence, which happens to be called strong
convergence, that requires that Tn f → T f as n → ∞ for every f ∈ H.
Finally, there is weak convergence (see also Exercise 20) that requires
(Tn f, g) → (T f, g) for every pair of vectors f, g ∈ H.
(a) Show by examples that weak convergence does not imply strong con-
vergence, nor does strong convergence imply convergence in the norm.
(b) Show that for any bounded operator T there is a sequence {Tn } of
bounded operators of finite rank so that Tn → T strongly as n → ∞.

Solution. (a) Let H = `2 (N). Let T be the zero operator and Tn = Rn

where R is the right shift operator; thus
Tn (a1 , a2 , . . . ) = (0, . . . , 0, a1 , a2 , . . . ).
Then for any fixed f = (a1 , a2 , . . . ) and g = (b1 , b2 , . . . ),
∞
X
hTn f, gi = ak bn+k = hf, Ln gi
k=1

where L is the left-shift operator defined by Lg = (b2 , b3 , . . . ). By the

Cauchy-Schwarz inequality,
|hTn f, gi| ≤ kf kkLn gk.
But kLn gk → 0 because |bk |2 converges and the tails of a convergent
P
series tend to zero. Hence hTn f, gi → 0 for all f, g ∈ H. Thus,
Tn → T weakly. However, Tn does not converge to T strongly; note
that kRf k = kf k for any f , so kTn f k = kRn f k = kf k by induction.
Hence Tn f 6→ T f = 0.
To show that strong convergence does not imply convergence in norm,
let Tn = Ln in the same space. Then for any f , kTn f k = kLn f k → 0
because the tails of a convergent series tend to zero. Hence Tn → T
strongly. However, kTn k = 1 because a unit vector (0, . . . , 0, 1, 0, . . . )
with more than n initial zeros is mapped to another unit vector. Hence
kTn − T k = kTn k 6→ 0.
62

(b) (We assume H is separable.) Let {ei } be an orthonormal basis for H.

We can write
X∞
T ei = cij ej
j=1
for each i = 1, 2, . . . . Define
n
X
Tn ei = cij ej
j=1

and extend linearly from the basis to the rest of the space (actually,
extend linearly to finite linear combinations of the basis, and then take
limits to get the rest of the space...) Clearly each TnP is of finite rank
∞
since its range is spanned by e1 , . . . , en . Now let f = i=1 ai ei . Then
∞ X
X ∞
Tf = ai cij ej
i=1 j=1

whereas
∞ X
X n
Tn f = ai cij ej
i=1 j=1
which is just the nth partial sum (in j) and hence converges to T f .
(This is where we use the fact that T is a bounded operator, since
absolute convergence allows us to rearrange these sums.) Hence Tn f →
T f weakly for all f ∈ H.

Exercise 22: An operator T is an isometry if kT f k = kf k for all f ∈ H.
(a) Show that if T is an isometry, then (T f, T g) = (f, g) for every f, g ∈ H.
Prove as a result that T ∗ T = I.
(b) If T is an isometry and T is surjective, then T is unitary and T T ∗ = I.
(c) Give an example of an isometry that is not unitary.
(d) Show that if T ∗ T is unitary then T is an isometry.
Solution. (a) By the polarization identity,
kT f + T gk2 − kT f − T gk2 + ikT f + iT gk2 − ikT f − iT gk2
hT f, T gi =
4
kT (f + g)k2 − kT (f − g)k2 + ikT (f + ig)k2 − ikT (f − ig)k2
=
4
kf + gk2 − kf − gk2 + ikf + igk2 − ikf − igk2
=
4
= hf, gi.
This in turn implies
hf, T ∗ T gi = hf, Igi
for all f, g, so that T ∗ T = I.
(b) T preserves norms because it’s an isometry; it’s injective because norm-
preserving linear maps are always injective (since the kernel cannot
contain anything nonzero). Since it’s surjective as well, it’s a norm-
preserving linear bijection, which is by definition a unitary map. We
 63

know T ∗ T = I from part (a). Since T is bijective, it has a linear 2-

sided inverse, and the equation T ∗ T = 1 shows that T is this inverse.
Hence T T ∗ = I.
(c) The right-shift operator on `2 (N) is isometric, since
∞
X
k(0, a1 , a2 , . . . )k2 = |aj |2 = k(a1 , a2 , . . . )k2 .
j=1

However, this operator is not surjective, so it’s not unitary.

(d) First,
kT f k2 = hT f, T f i = hf, T ∗ T f i ≤ kf kkT ∗ T f k = kf k2
by the Cauchy-Schwarz inequality; hence kT f k ≤ kf k. Then,
kf k2 = kT ∗ T f k2 = hT ∗ T f, T ∗ T f i = hT f, T T ∗ T f i ≤ kT f kkT (T ∗ T f )k ≤ kT f kkT ∗ T f k = kT f kkf k
where we have applied the previous inequality with T ∗ T f in place of f .
Dividing by kf k yields kf k ≤ kT f k. Putting the inequalities together,
kT f k = kf k. Hence T is an isometry.


Exercise 23: Suppose {Tk } is a collection of bounded operators on a Hilbert

space H, with kTk k ≤ 1 for all k. Suppose also that
Tk Tj∗ = Tk∗ Tj = 0 for all k 6= j.
Pn
Let SN = k=−N Tk . Show that SN (f ) converges as N → ∞, for every
f ∈ H. If T (f ) denotes the limite, prove that kT k ≤ 1.
Exercise 24: Let {ek }∞ P 2 space H.
k=1 denote an orthonormal set in a Hilbert
If {ck }∞
k=1 is a sequence of positive real numbers such that ck < ∞, then
the set (∞ )
X
A= ak ek : |ak | ≤ ck
k=1
is compact in H.

Solution. This is a standard diagonalization argument. Let Σ be a sequence

of points in H. Consider the first components of the points in Σ, i.e. their
components with respect to e1 . This is a sequence of complex numbers
in the compact ball {z : |z| ≤ c1 }, so some subsequence converges to a
complex number in this ball. If we take the corresponding subsequence of
Σ, we obtain a subsequence S1 whose first components converge to some
number b1 with |b1 | ≤ c1 . Now consider the second components of the
points in S1 ; they form a sequence of complex numbers in the compact ball
{z : |z| ≤ c2 } and hence some subsequence converges to a complex number
b2 in this ball. Taking the corresponding subsequence S2 of S1 , we have a
sequence whose first and second components converge. Continuing, we may
inductively define sequences Sn for all n such that Sn+1 is a subsequence
of Sn , and the first n components of Sn converge. Finally, we define a
sequence S whose nth term is the nth term of Sn . This is a subsequence
of Σ, and for any n it is eventually a subsequence of Sn (i.e. the tail of S
is a subsequence of the tail of Sn ). Hence it converges in every component.
64

But this implies convergence to a point in the Hilbert space (since the sizes
of the tails are uniformly bounded), so we are done. 

Exercise 25: Suppose T is a bounded operator that is diagonal with respect

to a basis {φk }, with T φk = λk φk . Then T is compact if and only if λk → 0.

Solution. (From lecture) Suppose λk → 0. Let Tn be the nth truncation,

i.e. the operator that results when λk is replaced with 0 for k > n. Then
T − Tn is also a diagonal operator, with
kT − Tn k = sup|λk | → 0.
k>n

Since T can be uniformly approximated by operators of finite rank, it is

compact. Conversely, suppose that λk 6→ 0, i.e. lim sup |λn | > 0, so that
there is some subsequence λnj with |λnj | > 2δ for some real number δ > 0.
Then T φnj = λnj φnj and by orthonormality,
q δ
kT φnj − T φnk k = λ2nj + λ2nk > √ .
2
Since all the points of the sequence {T φnj } are uniformly bounded away
from each other, it can have no convergent subsequence. These points all
lie in T (B), so T (B) is not compact. 

Exercise 26: Suppose w is a measurable function on Rd with 0 < w(x) < ∞

for a.e. x, and K is a measurable function on R2d that satisfies:
(i)
Z
|K(x, y)|w(y)dy ≤ Aw(x) for almost every x ∈ Rd , and
Rd

(ii)
Z
|K(x, y)|w(x)dx ≤ Aw(y) for almost every y ∈ Rd .
Rd

Prove that the integral operator defined by

Z
T f (x) = K(x, y)f (y)dy, x ∈ Rd
Rd
2 d
R bounded on L (R ) with kTRk ≤ A. Note as a special case that if
is
|K(x, y)|dy ≤ A for all x, and |K(x, y)|dx ≤ A for all y, then kT k ≤ A.

Solution. First, note that

Z 2 Z   2
−1
p p  p p
|K(x, y)||f (y)|dy = |K(x, y)| w(y) |K(x, y)||f (y)| w(y)
Z  Z 
2 −1
≤ |K(x, y)|w(y)dy |K(x, y)||f (y)| w(y) dy
(a.e.)
Z
≤ Aw(x) |K(x, y)||f (y)|2 w(y)−1 dy
 65

by the Cauchy-Schwarz inequality. Thus, for f ∈ L2 ,

Z Z 2
kT f k2 = K(x, y)f (y)dy dx
Z Z 2
≤ |K(x, y)||f (y)|dy dx
Z Z
≤ Aw(x) |K(x, y)||f (y)|2 w(y)−1 dydx
Z Z
(T onelli) 2 −1
= A |f (y)| w(y) |K(x, y)|w(x)dxdy
Z
≤ A |f (y)|2 w(y)−1 Aw(y)dy
Z
2
=A |f (y)|2 dy

= A2 kf k2 .
Hence kT k ≤ A. 
Exercise 27: Prove that the operator
1 ∞ f (y)
Z
T f (x) = dy
π 0 x+y
is bounded on L2 (0, ∞) with norm kT k ≤ 1.
Exercise 28: Suppose H = L2 (B), where B is the unit ball in Rd . Let
K(x, y) be a measurable function on B × B that satisfies |K(x, y)| ≤ A|x −
y|−d+α for some α > 0, whenever x, y ∈ B. Define
Z
T f (x) = K(x, y)f (y)dy.
B
(a) Prove that T is a bounded operator on H.
(b) Prove that T is compact.
(c) Note that T is a Hilbert-Schmidt operator if and only if α > d/2.
Solution.
(a) Let Z
dz
C=
|z|d−α
z∈Rd :|z|≤2
which converges because the exponent is less than d. Then
Z Z Z
Ady dz
|K(x, y)|dy ≤ d−α
≤ A d−α
= AC
B |x − y| |z|≤2 |z|
so by problem 26 with w = 1, we have T bounded with kT k ≤ AC.
(b) As suggested, let
(
K(x, y) |x − y| ≥ n1
Kn (x, y) =
0 else
and Z
Tn f (x) = Kn (x, y)f (y)dy.
66

Then Tn is Hilbert-Schmidt (and therefore compact) since clearly Kn ∈

L2 (B×B) (Kn is, after all, bounded with compact support). Moreover,
Z Z
|Kn (x, y) − K(x, y)|dy ≤ A|x − y|−d+α = ACn
B |x−y|≤1/n

where we define
Z
dz
Cn = .
|z|d−α
z∈Rd :|z|≤1/n

1
Since |z|d−α ∈ L1 (Rd ), the absolute continuity of the integral implies
Cn → 0. By problem 26 again with w = 1, this implies kT − Tn k → 0.
Since Tn is compact, this implies that T is compact.
(c) This should actually say “T is guaranteed to be Hilbert-Schmidt if and
only if...” since K could be a lot less than the bound given. Anyhoo,
T necessarily Hilbert-Schmidt ⇔ A|x − y|−d+α ∈ L2 (B × B)
Z Z
⇔ A2 |x − y|−2d+2α < ∞
B B
⇔ −2d + 2α > −d
d
⇔α> .
2


Exercise 29: Let T be a compact operator on a Hilbert space H and assume

λ 6= 0.
(a) Show that the range of λI − T defined by
{g ∈ H : g = (λI − T )f for some f ∈ H}
is closed.
(b) Show by example that this may fail when λ = 0.
(c) Show that the range of λI − T is all of H if and only if the null space
of λ̄I − T ∗ is trivial.
Exercise 30: Let H = L2 ([−π, π]) with [−π, π] identified as the unit circle.
Fix a bounded sequence {λn }∞ n=−∞ of complex numbers, and define an
operator T f by
∞
X ∞
X
inx
T f (x) ∼ λn an e whenever f (x) ∼ an einx .
n=−∞ n=−∞

Such an operator is called a Fourier multiplication operator, and the

sequence {λn } is called the multiplier sequence.
(a) Show that T is a bounded operator on H and kT k = sup |λn |.
(b) Verify that T commutes with translations, that is, if we define τh (x) =
f (x − h) then
T ◦ τh = τh ◦ T for every h ∈ R.
(c) Conversely, prove that if T is any bounded operator on H that com-
mutes with translations, then T is a Fourier multiplier operator.
 67

Exercise 34: Let K be a Hilbert-Schmidt kernel which is real and symmet-

ric. Then, as we saw, the operator T whose kernel is K is compact and
symmetric. Let {φk (x)} be the eigenvectors (with eigenvalues λk ) that
diagonalize T . Then
|λk |2 < P
P
(a) ∞.
(b) K(x, y) ∼ λk φk (x)φk (y) is the expansion of K in the basis {phik (x)φk (y)}.
(c) Suppose T is a compact operator which P is2 symmetric. Then T is of
Hilbert-Schmidt type if and only if |λn | < ∞, where {λn } are the
eigenvalues of T counted according to their multiplicities.
Exercise 35: Let H be a Hilbert space. Prove the following variants of the
spectral theorem.
(a) If T1 and T2 are two linear symmetric and compact operators on H
that commute, show that they can be diagonalized simultaneously. In
other words, there exists an orthonormal basis for H which consists of
eigenvectors for both T1 and T2 .
(b) A linear operator on H is normal if T T ∗ = T ∗ T . Prove that if T is
normal and compact, then T can be diagonalized.
(c) If U is unitary, and U = λI − T where T is compact, then U can be
diagonalized.

Solution.
(a) We can pretty much copy the proof verbatim with “eigenvector” re-
placed by “common eigenvector”. Let S be the closure of the subspace
of H spanned by all common eigenvectors of T1 and T2 . We want to
show S = H. Suppose not; then H = S ⊕ S ⊥ with S ⊥ nonempty. If we
can show S ⊥ contains a common eigenvector of T1 and T2 , we have a
contradiction. Note that T1 S ⊂ S, which in turn implies T1 S ⊥ ⊂ S ⊥
since
g ∈ S ⊥ ⇒ hT g, f i = hg, T f i = 0
for all f ∈ S. Similarly, T2 S ⊥ ⊂ S ⊥ . Now by the theorem for one
operator, T1 must have an eigenvector in S ⊥ with some eigenvalue λ.
Let Eλ be the eigenspace of λ (as a subspace of S ⊥ ). Then for any
x ∈ Eλ ,
T1 (T2 x) = T2 (T1 x) = T2 (λx) = λ(T2 x)
so T2 x ∈ Eλ as well. Since T2 fixes Eλ , it has at least one eigenvector in
Eλ . This eigenvector is a common eigenvector of T1 and T2 , providing
us with our contradiction.
(b) This follows from part (a). Write
T + T∗ T − T∗
T = +i .
2 2i
T +T ∗ T −T ∗
By a trivial calculation, both 2 and 2i are self-adjoint. More-
over, since T is normal,
(T + T ∗ )(T − T ∗ ) = T 2 + T ∗ T − T T ∗ − T ∗2 = T 2 − T ∗2 = (T − T ∗ )(T + T ∗ )
so they commute as well. Hence, there exists an ONB of common
∗ ∗
eigenvectors of T +T
2 and T −T
2i . Any such common eigenvector is an
68

eigenvector of T , since
T + T∗ T − T∗
x = λx and x = λ0 x ⇒ T x = (λ + iλ0 )x.
2 2i
(c)


Chapter 4.8, Page 202

Problem 1: Let H be an infinite-dimensional Hilbert space. There exists a
linear functional ` defined on H that is not bounded.

Solution. It is a well-known fact from linear algebra that every vector space
has a basis. This can be proved using Zorn’s lemma: linearly independent
sets are partially ordered by inclusion, and every chain has an upper bound
by union, so there exists a maximal linearly independent set, which is by
definition a basis. Applying this to our Hilbert space, we obtain an (alge-
braic) basis, i.e. one for which every vector is a finite linear combination
of basis elements. Let {en } be a countable subset of our algebraic basis.
Define `(en ) = nken k and `(f ) = 0 for f in our basis but f 6= en for any
n. We can then extend ` to the whole space in a well-defined manner, but
clearly ` is not bounded since |`(en )| = nken k. 

Problem 2: The following is an example of a non-separable Hilbert space.

We consider the collection of exponentials {eiλx } on R, where λ ranges over
the real numbers. Let H0 denote the space of finite linear combinations of
these exponentials. For f, g ∈ H0 , we define the inner product as
Z T
1
(f, g) = lim f (x)g(x)dx.
T →∞ 2T −T

(a) Show that this limit exists, and

N
X
(f, g) = aλk bλ¯k
k=1
PN PN
if f (x) = k=1 aλk eiλk x and g(x) = k=1 bλk eiλk x .
(b) With this inner product H0 is a pre-Hilbert space. Notice that kf k ≤
supx |f (x)|, if f ∈ H0 , where kf k denotes the norm hf, f i1/2 . Let H
be the completion of H0 . Then H is not separable because eiλx and
0
eiλ x are orthonormal if λ 6= λ0 . A continuous function F defined on R
is called almost periodic if it is the uniform limit (on R) of elements
in H0 . Such functions can be identified with (certain) elements in the
completion H: We have H0 ⊂ AP ⊂ H, where AP denotes the almost
periodic functions.
(c) A continuous function F is in AP if for ever  > 0 we can find a length
L = L such that any interval I ⊂ R of length L contains an “almost
period” τ satisfying
sup|F (x + τ ) − F (x)| < .
x
 69

(d) An equivalent characterization is that F is in AP if and only if every

sequence F (x + hn ) of translates of F contains a subsequence that
converges uniformly.
Problem 7: Show that the identity operator on L2 (Rd ) cannot be given as
an (absolutely) convergent integral operator. More precisely, if K(x, y) is a
measurable function onRRd ×Rd with the property that for each f ∈ L2 (Rd ),
the integral T (f )(x) = Rd K(x, y)f (y)dy converges for almost every x, then
T (f ) 6= f for some f .
Solution. Suppose such a K exists. Let B1 and B2 be disjoint balls in Rd .
We will show that K = 0 a.e. in B1 × B2 . Suppose not; then there is a
“rectangle” E1 × E2 with E1 ⊂ B1 and E2 ⊂ B2 sets of positive measure,
on which K(x, y) > 0 or K(x, y) < 0; WLOG, K(x, y) > 0. (If K is allowed
complex values, we can change this condition to Re(K) > 0.) Let f = χE2 .
Then for almost all x ∈ E1 ,
Z
0 = f (x) = T f (x) = K(x, y)dy.
E2
But the integral of a positive function over a set of positive measure is
nonzero, so we have a contradiction. Thus, K(x, y) = 0 a.e. in B1 × B2 .
Now if we let ∆ = {(x, x) : x ∈ Rd } ⊂ R2d be the “diagonal”, we can cover
R2d \ ∆ with product sets B1 × B2 . (One way to see this is topological:
products of balls form a basis for the product topology on Rd × Rd , and
∆ is closed, so its complement is open.) This implies K(x, y) = 0 a.e. on
R2d \ ∆, and therefore a.e. on R2d . But then T f = 0 for all f , so any
nonzero f will have T f 6= f . 
Problem 8: Suppose {Tk } is a collection of bounded operators on a Hilbert
space H. Assume that
kTk Tj∗ k ≤ a2k−j
and kTk∗ Tj k ≤ a2k−j ,
P∞
for positive constants {an } with the property that n=−∞ an = A < ∞.
PN
Then SN (f ) converges as N → ∞, for every f ∈ H, with SN = −N Tk .
Moreover, T = limN →∞ SN satisfies kT k ≤ A.
Solution. For any integers N and n,
n n−1
kSN k2 = k(S ∗ S)2 k
N
X N
X N
X N
X
= ··· Tj∗1 Tk1 · · · Tj∗2n−1 Tk2n−1
j1 =−N k1 =−N j2n−1 =−N k2n−1 =−N
N
X N
X N
X N
X
≤ ··· kTj∗1 Tk1 · · · Tj∗2n−1 Tk2n−1 k.
j1 =−N k1 =−N j2n−1 =−N k2n−1 =−N

Now since kTm k ≤ a0 for any m,

70


Problem 9: A discussion of a class of regular Sturm-Liouville operators fol-
lows. Other special examples are given in the problems below.
Suppose [a, b] is a bounded interval, and L is defined on functions f that
are twice continuously differentiable in [a, b] (we write f ∈ C 2 ([a, b]) by
d2 f
L(f )(x) = − q(x)f (x).
dx2
Here the function q is continuous and real-valued on [a, b], and we assume
for simplicity that q is non-negative. We say that φ ∈ C 2 ([a, b]) is an
eigenfunction of L with eigenvalue µ if L(φ) = µφ, under the assumption
that φ satisfies the boundary conditions φ(a) = φ(b) = 0. Then one can
show:
(a) The eigenvalues µ are strictly negative, and the eigenspace correspond-
ing to each eigenvalue is one-dimensional.
(b) Eigenvectors corresponding to distinct eigenvalues are orthogonal in
L2 ([a, b]).
(c) Let K(x, y) be the “Green’s kernel” defined as follows. Choose φ− (x)
to be a solution of L(φ− ) = 0, with φ− (a) = 0 but φ0− (a) 6= 0. Simi-
larly, choose φ+ (x) to be a solution of L(φ+ ) = 0 with φ+ (b) = 0 but
φ0+ (b) 6= 0. Let w = φ0+ (x)φ− (x) − φ0− (x)φ+ (x), be the “Wronskian”
of these solutions, and note that w is a non-zero constant.
Set (
φ− (x)φ+ (y)
w a ≤ x ≤ y ≤ b,
K(x, y) = φ+ (x)φ − (y)
w a ≤ y ≤ x ≤ b.
Then the operator T defined by
Z b
T (f )(x) = K(x, y)f (y)dy
a
is a Hilbert-Schmidt operator, and hence compact. It is also sym-
metric. Moreover, whenever f is continuous on [a, b], T f is of class
C 2 ([a, b]) and
L(T f ) = f.
(d) As a result, each eigenvector of T (with eigenvalue λ) is an eigen-
vector of L (with eigenvalue µ = 1/λ). Hence Theorem 6.2 proves
the completeness of the orthonormal set arising from normalizing the
eigenvectors of L.
Solution.
(a) Let φ be an eigenfunction of L with eigenvalue µ. Then
φ00 = (q + µ)φ ⇒ φφ00 = (q + µ)φ2 .
Integrating by parts from a to b, we have
Z Z
φφ0 |ba − (φ0 )2 = (q + µ)φ2 .

Since φ(a) = φ(b) = 0, this reduces to

Z Z
0 2
− (φ ) = (q + µ)φ2 .
 71

Now if φ is not almost everywhere zero, the LHS is strictly negative.

But the integrand on the RHS is everywhere nonnegative unless µ is
strictly negative. Hence the eigenvalues of L are all strictly negative.
Now suppose µ is an eigenvalue of L with eigenfunctions φ1 and φ2 .
Then
(µ + q)φ1 φ2 = φ1 φ002 = φ2 φ001
⇒ φ01 φ02 + φ1 φ002 = φ01 φ02 + φ001 φ2
⇒ (φ1 φ02 )0 = (φ01 φ2 )0
⇒ φ1 φ02 = φ01 φ2 + C.
Plugging in a, we see that C = 0 because φ1 (a) = φ2 (a) = 0. Hence
φ1 φ02 − φ01 φ2 = 0. Since the Wronskian of these solutions is zero, they
are linearly dependent. So the eigenspace can only be one-dimensional.
(b) Suppose φ1 and φ2 are eigenfunctions with eigenvalues µ1 and µ2 re-
spectively. Then
Z Z
µ1 φψ = (φ00 − qφ)ψ
Z Z
00
= φ ψ − qφψ
Z Z
= φ0 ψ|ba − φ0 ψ 0 − qφψ
Z Z
= − φ0 ψ 0 − qφψ
Z Z
= φψ 0 |ba − φ0 ψ 0 − qφψ
Z Z
= ψ 00 φ − qφψ
Z
= (ψ 00 − qψ)φ
Z
= µ2 ψφ.
R
If µ1 6= µ2 , this implies φ1 φ2 = 0.
(c) The Wronskian is a constant because
w0 = φ00+ φ− − φ00− φ+ = qφ+ φ− − qφ− φ+ = 0;
it is nonzero because plugging in at a yields −φ0− (a)φ+ (a). We already
know φ0− (a) 6= 0, and φ+ (a) cannot be zero because then φ+ would be
an eigenfunction with eigenvalue
R R 0, 2contradicting part (a). To show T
is Hilbert-Schmidt, consider |K |. We will treat this on the region
R = {a ≤ y ≤ x ≤ b}; the other half is symmetric. Then
|φ+ (x)φ− (y)|2
ZZ ZZ
|K(x, y)|2 dxdy = dxdy.
R R w2
Now w is a nonzero constant, as we saw above; φ+ and φ− are both
continuous on a compact set and hence bounded. Thus, the integrand
is bounded, and the region of integration is compact, so the integral
72

is finite. So T is Hilbert-Schmidt. The symmetry of T is immedi-

ately evident from its definition. Now suppose f ∈ C([a, b]). Then
T f ∈ C 2 ([a, b]) because K ∈ C 2 ([a, b]2 ) and the second partials of K
are bounded so that one can differentiate T f under the integral sign.
Finally,
Z b
1 x 1 b
Z Z
T f (x) = K(x, y)f (y)dy = φ+ (x)φ− (y)f (y)dy + φ− (x)φ+ (y)f (y)dy
a w a w x
φ+ (x) x φ− (x) b
Z Z
= φ− (y)f (y)dy + φ+ (y)f (y)dy,
w a w x

so
φ0 (x) x
φ0 (x) b
Z Z
φ+ (x) φ− (x)
(T f ) (x) = +
0
φ− (y)f (y)dy + φ− (x)f (x) + − φ+ (y)f (y)dy − φ+ (x)f (x)
w a w w x w
0 Z x 0 Z b
φ (x) φ (x)
= + φ− (y)f (y)dy + − φ+ (y)f (y)dy
w a w x

and
φ00 (x) x
φ0+ (x) φ00− (x) b φ0 (x)
Z Z
(T f ) (x) = +
00
φ− (y)f (y)dy + φ− (x)f (x) + φ+ (y)f (y)dy − − φ+ (x)f (x)
w a w w x w
w q(x)φ+ (x) x q(x)φ− (x) b
Z Z
= f (x) + φ− (y)f (y)dy + φ+ (y)f (y)dy
w w a w x
Z b
= f (x) + q(x) K(x, y)f (y)dy
a
= f (x) + q(x)(T f )(x)

so L(T f ) = (T f )00 − q(T f ) = f .

(d) This is more just an observation in the problem statement than some-
thing for me to do.


Chapter 5.5, Page 253

Exercise 1: Suppose f ∈ L2 (RRd ) and k ∈ L1 (Rd ).
(a) Show that (f ∗ k)(x) = f (x − y)k(y)dy converges for a.e. x.
(b) Prove that kf ∗ kk2 ≤ kf k2 kkk1 .
\
(c) Establish (f ∗ k)(ξ) = k̂(ξ)fˆ(ξ) for a.e. ξ.
(d) The operator T f = f ∗ k is a Fourier multiplication operator with
multiplier m(ξ) = k̂(ξ).

Solution.
(a) This will follow from part (b) because an L2 function must R be fi-
nite almost everywhere (which will prove a.e. convergence of |f (x −
y)||k(y)|dy), and absolutely convergent integrals are convergent.
(b) Just so I have it for future reference, why don’t I prove the Lp version
of this. Suppose f ∈ Lp with 1 < p < ∞ and k ∈ L1 . Let q be the
 73

conjugate exponent of p. Then

Z Z p
kf ∗ kkpp = f (x − y)k(y)dy dx
Z Z p
1/p 1/q
≤ |f (x − y)||k(y)| |k(y)| dy dx
Z  p
≤ kf (x − y)k(y)1/p kp kk(y)1/q kq dx
Z Z  Z p/q
p
= |f (x − y)| |k(y)|dy |k(y)|dy dx
p/q
= kkk1 kf p ∗ kk1
p/q
≤ kkk1 kkk1 kf p k1
= kkkp1 kf kpp .

Here we have used Hölder’s inequality on |f ||k|1/p ∈ Lp and |k|1/q ∈

Lq , as well as the bound for the L1 norm of a convolution of L1 func-
tions. In the case p = q = 2, Hölder’s inequality reduces to Cauchy-
Schwarz.
(c) If f ∈ L1 ∩ L2 , we already know this from our theory of Fourier
transforms on L1 . Otherwise, since L1 ∩ L2 is dense in L2 , we may
L2
take a sequence fn ∈ L1 ∩ L2 with fn → f . Then
Z Z
∗ k(ξ) − f\
f[ n ∗ k(ξ) = e−2πiξẋ (f (x − y) − fn (x − y)) k(y)dydx
Z Z
≤ |f (x − y) − fn (x − y)| |k(y)|dydx
L2
≤ kf − fn k2 kkk1 → 0
so
∗ k(ξ) = lim f\
f[ ˆ ˆ ˆ
n ∗ k(ξ) = lim fn (ξ)k̂(ξ) = k̂(ξ) lim fn (ξ) = k̂(ξ)f (ξ).

(d) This is just the definition of a Fourier multiplication operator applied

to part (c).


Exercise 3: Let F (z) be a bounded holomorphic function in the half-plane.

Show in two ways that limy→0 F (x + iy) exists for a.e. x.
(a) By using the fact that F (z)/(z
 + i) is in H 2 (R2+ ).
(b) By noting that G(z) = F i 1−z1+z is a bounded holomorphic function
in the unit disc, and using Exercise 17 in the previous chapter.

Solution.
(a) Since |F (z)| ≤ M for some M ,
F (x + iy) M
≤√
x + i(y + 1) x2 + 1
74

so
∞ 2 ∞
M2
Z Z
F (x + iy)
dx ≤ dx = M 2 π.
−∞ x + i(y + 1) −∞ x2 + 1
F (z)
Hence z+i ∈ H 2 (R2+ ). This implies that
F (x + iy)
lim
y&0 x + i(1 + y)

exists a.e., which in turn implies that lim F (x + iy) exists a.e.
(b) I assume that I can take for granted that z 7→ i 1−z 1+z is a conformal
mapping of the unit disc into the upper half plane, since we did this
on a previous homework. Then define G(w) = F (i 1−w 1+w ) which is a
bounded holomorphic function on D. It now suffices to show that w
approaches the unit circle non-tangentially as y = Re(z) → 0, where
w is now given by the inverse mapping
−x + (1 − y)i
w= .
x + (1 + y)i
Then
x2 + (1 − y)2
|w|2 =
x2 + (1 + y)2
by straightforward arithmetic. Now if w were approaching the unit
2
circle in a tangential manner, we would have d|w|
dy |y=0 = 0. However,

d|w|2 4(y 2 − x2 − 1)
= 2
dy (x + (1 − y)2
which is nonzero at y = 0.


Exercise 4: Consider F (z) = ei/z /(z + i) in the upper half-plane. Note that
F (x + iy) ∈ L2 (R), for each y > 0 and y = 0. Observe also that F (z) → 0
as |z| → ∞. However, F ∈ / H 2 (R2+ ). Why?

Solution. For any fixed y > 0,

2
ei/(x+iy) e1/y
≤
x + i(1 + y) x2 + 1
which is integrable, so F (x + iy) ∈ L2 (R). For y = 0,
2
ei/x 1 1
= = 2
x+i |x + i|2 x +1
which is again integrable. Also, as |z| → ∞, the numerator approaches 1 in
magnitude while the denominator becomes infinite. However, F ∈ / H 2 , as is
1/y
suggested by the fact that our bound includes an e term, which blows up.
The problem, of course, is that F is not bounded in the upper half plane;
it has an essential singularity at 0, and Picard’s theorem tells us that it
takes on every complex value (except possibly 1) in every neighborhood of
the origin. 
 75

Exercise 6: Suppose Ω is an open set in C = R2 , and let H be the subspace

of L2 (Ω) consisting of holomorphic functions on Ω. Show that H is a closed
subspace of L2 (Ω), and hence is a Hilbert space with inner product
Z
(f, g) = f (z)ḡ(z)dxdy, where z = x + iy.
Ω

Solution. For any f ∈ H, f 2 ∈ H as well, since the square of an analytic

function is analytic. Now by the mean value property, for any z ∈ Ω and
r ≤ d(z, Ωc ),
ZZ
1
f (z)2 = 2 f (ζ)2 dA,
πr
|ζ−z|≤r

whence ZZ
1 1
|f (z)|2 ≤ |f (ζ)|2 dA ≤ kf k2 .
πr2 πr2
|ζ−z|≤r

Now on any compact K ⊂ Ω, there is a minimum value r0 > 0 of d(z, Ωc ) for

z ∈ K. (This is because the distance between a compact set and a √closed
set always attains a nonzero minimum.) Then we have |f (z)| ≤ r0π kf k
for all z ∈ K and f ∈ L2 (Ω). So if {fn } is a Cauchy sequence in L2 (Ω),
then kfm − fn k → 0, whence |fm − fn | → 0 as well. Thus, {fn } converges
uniformly on any compact subset of Ω. Now it is a theorem in complex
analysis that the uniform limit of analytic functions is analytic; this may be
proved, for example, by using the M L estimate to show that the integral of
the limit around any contour is zero, and then applying Morera’s theorem.
(See e.g. Gamelin p. 136.) This theorem works on any domain, e.g. the
interior of any compact disc contained in Ω. This allows us to prove that
the limit of {fn } is analytic at each point in Ω, so it’s analytic on Ω. 

Exercise 7: Following up on the previous exercise, prove:

(a) If {φn }∞
n=0 is an orthonormal basis of H, then
∞ 2
X c2
|φn (z)|2 ≤ for z ∈ Ω.
n=0
d(z, Ωc )

(b) The sum

∞
X
B(z, w) = φn (z)φ¯n (w)
n=0

converges absolutely for (z, w) ∈ Ω × Ω, and is independent of the

choice of the orthonormal basis {φn } of H.
(c) To prove (b) it is useful to characterize the function B(z, w), called
the Bergman kernel, by the following property. Let T be the linear
transformation on L2 (Ω) defined by
Z
Tf = B(z, w)f (w)dudv, w = u + iv.
Ω

Then T is the orthogonal projection of L2 (Ω) to H.

76

Suppose that Ω is the unit disc. Then f ∈ H exactly when f (z) =

(d) P
∞ n
n=0 an z , with
∞
X
|an |2 (n + 1)−1 < ∞.
n=0
n
Also, the sequence { z (n+1) ∞
π 1/2
}n=0 is an orthonormal basis of H. More-
over, in this case
1
B(z, w) = .
π(1 − z w̄)2
Solution.
(a) First we prove a lemma, whose relevance was so kindly pointed out by
Prof. Garnett:
Lemma 3. Let {bn }∞n=0 be a sequence of complex numbers. Then
v
u∞ ∞
uX X
t |bn |2 = P sup an bn .
n=0 |an |2 ≤1 n=0

Proof. If |bn |2 < ∞ this follows from the Cauchy-Schwarz inequality

P
applied to `2 (N), where equality is achieved when {an } is the unit
vector inPthe same direction (actually, the conjugate) as {bn }. Now
suppose |bn |2 = ∞. Then for any N , define the truncated sequence
(N ) (N )
b̃ = {b̃n } by
(
bn n ≤ N
b̃(N
n
)
=
0 else.
(N )
Then b̃(N ) ∈ `2 , so if a(N ) = {an }∞ n=0 is the unit vector in the
conjugate direction of b̃(N ) , we have
∞
X ∞
X
a(N )
n bn = an(N ) b̃n(N ) = kb̃(N ) k.
n=0 n=0

Since this goes to infinity as N → ∞, we have

v
u∞
X uX
P
sup an bn = ∞ = t |bn |2 .
kan k2 ≤1 n=0


|an |2 ≤
P
Returning to the problem at hand, for any sequence an with
1,
∞
X
g(z) = an φn (z)
n=0
is a unit vector in H. Applying problem 6, we have at any fixed z ∈ Ω
that
∞ √ √
X π π
an φn (z) = |g(z)| ≤ c
kgk = .
n=0
d(z, Ω ) d(z, Ωc )
 77

Applying the lemma with bn = φn (z), we have

∞ ∞
!2
X X π
|φn (z)|2 = P sup an φn (z) ≤ .
n=0
2
|an | ≤1 n=0 d(z, Ωc )2

(b) The absolute convergence of this sum follows from part (a) and the
Cauchy-Schwarz inequality: For fixed values of z and w, {|φn (z)|} and
{φ¯n (w)} are vectors in `2 , so by the Cauchy-Schwarz inequality,
qX qX
φn (z)φn¯(w) ≤
X
|φn (z)|2 |φ¯n (w)|2 < ∞.

To prove that the sum is independent of the choice of basis, we use

part (c). Because integration against this sum is projection onto H,
and there is only one projection map, any two such sums must be
equal almost everywhere. I’m not 100% sure how to go about show-
ing they are in fact equal everywhere. Certainly B(z, w) is analytic
in z and analytic in w (with either variable fixed, it’s in H as a func-
tion of the other variable). However, I don’t know anything about
functions of several complex variables; is a function that’s analytic in
each variable separately necessarily analytic? Or, more to the point,
continuous? Assuming so, continuity plus equality almost everywhere
implies equality. Of course, one could say that since H is being viewed
as a subspace of L2 , a.e. equality is all we need for the functions to be
the same point in the Hilbert space.
(c) Since {φn } is an ONB for the closed subspace H, we can extend it
to a basis for all of L2 by complementing it with another set {ψk } of
orthonormal vectors. For z ∈ Ω, define Bz (w) = B(z, w). Then
Z
T f (z) = Bz (w)f (w)dw = hf, Bz i.
Ω
P∞ P∞
We can write f in our ONB as f (w) = n=0 an φn (w)+ k=0 bk ψk (w)
whence

T f (z) = hf, Bz i
* ∞ ∞
! ∞ +
X X X
= an φn + bk ψk ,  φ¯j (z)φj 
n=0 k=0 j=0
X X
= an φj (z)hφn , φj i + bk φj (z)hφk , φj i
j,n k,j
X
= an φn (z).
n

This is the formula for projection onto a closed subspace–T erases all
the components nin the
q orthogonal
o complement.
n+2
(d) The set {φn } = z n π is orthonormal since

1 2π
r2n+2
Z Z Z 1
n+1 2 n+1
hφn , φn i = |z n|dA = r2 nrdrdθ = 2(n + 1) =1
π D π r=0 θ=0 2n + 2 0
78

and
p
(m + 1)(n + 1) 1
Z Z 2π
hφn , φi = rn+m e2πi(n−m) rdrdθ = 0
π r=0 θ=0

for m 6= n. Now since every analytic function hasP a power series ex-
pansion, any analytic function can be written as bn φn . This proves
that {φn } is a basis for H, and also gives us the condition for an
P |an |2
analytic function to be in L2 : |bn |2 < ∞ ⇔
P
q n+1 < ∞, since
π
bn = an n+1 .
To obtain an expression for B(z, w), we first note that for any complex
number ζ with |ζ| < 1,
∞
1 X
= (n + 1)ζ n .
(1 − ζ)2 n=0

This may be obtained by differentiating the series ζ1 = ζ n termwise,

P

or by squaring it and collecting like terms. Both are justified by the

uniform absolute convergence of this series on compact subdisks of the
unit disk. Then
∞ r r ∞
X n+1 n n+1 n 1X π
B(z, w) = z w̄ = (n + 1)(z w̄)n = .
n=0
π π π n=0
(1 − z w̄)2

Exercise 8: Continuing with Exercise 6, suppose Ω is the upper half-plane

R2+ . Then every f ∈ H has the representation
√ Z ∞
f (z) = 4π fˆ0 (ξ)e2πiξz dξ, z ∈ R2+ ,
0
R∞
where 0
|fˆ0 (ξ)|2 dξ ˆ
ξ < ∞. Moreover, the mapping f0 → f given by this
formula is a unitary mapping from L2 ((0, ∞), dξ
ξ ) to H.

Solution. Following the proof of Theorem 2.1 on page 214, we define fˆy (ξ)
to be the Fourier transform of the L2 function f (x+iy). (We know f (x+iy)
is an L2 function of x for almost all y since
Z Z 
2 2
kf k2 = |f (x + iy)| dx dy
R
so that |f (x + iy)|2 dx is an integrable function of y, and therefore finite
almost everywhere. Then we can show that fˆy (ξ)e2πyξ is independent of y
using exactly the same proof in the book. (Our proof of the boundedness
of f on closed half-planes changes slightly: we now have
Z
1 1
|f (ζ)|2 = 2 |f (ζ + z)|2 dxdy ≤ 2 kf k22 .
δ |z|<δ δ
Other than that the proof requires no modification.) Having established
this, we can then define fˆ0 (ξ) to be the function that equals fˆy (ξ)e2πyξ for
 79

almost all y. The Plancherel formula then gives us

Z ∞ Z ∞
|f (x + iy)|2 dx = |fˆ0 (ξ)|2 e−4πyξ dξ.
−∞ −∞

This tells us that fˆ0 (ξ) = 0 for a.a. ξ < 0 (since the integral in ξ is infinite
for ξ < 0), and also gives us the relation
Z ∞Z ∞
kf k22 = |f (x + iy)|2 dxdy
−∞ −∞
Z ∞Z ∞
= |fˆ0 (ξ)|2 e−4πyξ dξdy
−∞ 0
Z ∞ Z ∞
Tonelli
= |fˆ0 (ξ)|2 e−4πyξ dydξ
0 −∞
Z ∞
1
= |fˆ0 (ξ)| 2
dξ
0 4πξ
This tells us that kf k = k √14π f0 kL2 ((0,∞),dξ/ξ) . We also have by Fourier
√ R∞
inversion that f (z) = 4π 0 √14π fˆ0 (ξ)e2πiz dξ. If we replace fˆ0 by √14π fˆ0 ,
we will have a unitary map fˆ0 → f , and
√ Z ∞
f (z) = 4π fˆ0 (ξ)e2πiξz dξ.
0

Exercise 9: Let H be the Hilbert transform. Verify that
(a) H ∗ = −H, H 2 = −I, and H is unitary.
(b) If τh denotes the translation operator, τh (f )(x) = f (x − h), then H
commutes with τh , τh H = Hτh .
(c) if δa denotes the dilation operator, δa (f )(x) = f (ax) with a > 0, then
H commutes with δa , δa H = Hδa .
Solution.
(a) Since the projection P and the identity I are both self-adjoint, 2P − I
is self-adjoint, so H = −i(2P − I) is skew-adjoint.
(b) Since H is a linear combination of I and P , it suffices to verify that
both of these commute with τh . For I this is trivial. For P , we have
2πih ˆ
P\
(τh f )(ξ) = χ(ξ)τd
h f (ξ) = χ(ξ)e f (ξ)
and
τ\
h P (f )(ξ) = e P f (ξ) = e2πih χ(ξ)fˆ(ξ).
2πih c

Since the Fourier transform on L2 is invertible, this implies τh P f =

P τh f , so P commutes with τh .
(c) Again, it suffices to verify that P commutes with dilations.
δ\ ˆ ˆ
a P (f )(ξ) = aP f (aξ) = aχ(aξ)f (aξ) = aχ(ξ)f (aξ)
c
where χ(aξ) = χ(ξ) because a > 0. Similarly,
P[
δa f (ξ) = χ(ξ)δd ˆ
a f (ξ) = χ(ξ)af (aξ).

Hence P commutes with dilations.

80

Exercise 15: Suppose f ∈ L2 (Rd ). Prove that there exists g ∈ L2 (Rd ) such
that  α
∂
f (x) = g(x)
∂x
in the weak sense, if and only if
(2πiξ)α fˆ(ξ) = ĝ(ξ) ∈ L2 (Rd ).
∂ α ∂ α
Then L∗ = (−1)|α|



Solution. (Help from Kenny Maples.) Let L = ∂x . ∂x .
Note in particular that

\  α
∂
∗ ψ(ξ) = (−1)|α|
Ld ψ(ξ) = (−1)|α| (2πiξ)α ψ̂(ξ) = (2πiξ)α ψ̂(ξ).
∂x

Now suppose ĝ = fˆ(ξ)(2πiξ)α ∈ L2 . Define g ∈ L2 as the inverse Fourier

transform of ĝ. Using Plancherel’s identity, for any ψ ∈ C0∞ we have

hg, ψi = hĝ, ψ̂i

Z
= ĝ(ξ)ψ̂(ξ)dξ
Z
= fˆ(ξ)(2πiξ)α ψ̂(ξ)dξ
Z
= fˆ(ξ)Ld ∗ ψ(ξ)dξ

= hfˆ, Ld
∗ ψi

= hf, L∗ ψi.
Hence g = Lf weakly.
Conversely, suppose there exists g ∈ L2 such that g = Lf weakly. Using
Plancherel again,
Z
ĝ(ξ)ψ̂(ξ)dξ = hĝ, ψ̂i

= hg, ψi
= hf, L∗ ψi
= hfˆ, Ld
∗ ψi
Z
= fˆ(ξ)(2πiξ)α ψ̂(ξ)dξ.

Since this is true for all ψ ∈ C0∞ , we must have ĝ(ξ) = fˆ(ξ)(2πiξ)α a.e.
Since g ∈ L2 , ĝ ∈ L2 by Plancherel, so fˆ(ξ)(2πiξ)α = ĝ(ξ) ∈ L2 . 

Chapter 5.6, Page 259

Problem 6: This problem provides an example of the contrast between anal-
ysis on L1 (Rd ) and L2 (Rd ).
 81

Recall that if f is locally integrable on Rd , the maximal function f ∗ is

defined by Z
1
f ∗ (x) = sup |f (y)|dx,
x∈B m(B) B
where the supremum is taken over all balls containing the point x.
Complete the following outline to prove that there exists a constant C
so that
kf ∗ k2 ≤ Ckf k2 .
In other words, the map that takes f to f ∗ (although not linear) is bounded
on L2 (Rd ). This differs notably from the situation in L1 (Rd ), as we observed
in Chapter 3.
(a) For each α > 0, prove that if f ∈ L2 (Rd ), then
Z
2A
m({x : f ∗ (x) > α}) ≤ |f (x)|dx.
α |f |>α/2
Here, A = 3d will do.
(b) Show that
Z Z ∞
|f ∗ (x)|2 dx = 2 αm(Eα )dα,
Rd 0
where Eα = {x : f ∗ (x) > α}.
(c) Prove that kf ∗ k2 ≤ Ckf k2 .
Solution.
α 2
(a) Let Gα = {x : |f (x)| > 2 }. Then 1 ≤ α |f | on Gα , so
Z Z
2 2
|f (y)|dy ≤ |f (y)|2 dy ≤ kf k2 < ∞.
Gα Gα α α
∗
Now let Eα = {x : f (x) > α}. For any x ∈ Eα , ∃Bx with x ∈ Bx and
Z  Z  Z
1 1 α 2
m(Bx ) < |f (y)|dy < m(Bx ) + |f (y)|dy ⇒ m(Bx ) < |f (y)|dy.
α Bx α 2 Gα ∩Bx α Gα ∩Bx
Here we have broken up the integral into the integral over the portion
of Bx where |f | ≤ α2 and the region where |f | > α2 and bounded each
portion. Now let K be any compact subset of Eα ; then K is covered
by finitely many balls Bx1 , . . . , BxN . By the Covering Lemma, there
exists a subcollection Bxn1 , . . . , BxnM such that
N
! M
[ X
m Bi ≤ 3d m(Bxnj ).
i=1 j=1

Then
M M Z
2 · 3d X 2 · 3d
X Z
m(K) ≤ 3d m(Bxnj ) ≤ |f (y)|dy ≤ |f (y)|dy.
j=1
α j=1 Gα ∩Bxn α Gα
j

By the regularity of Lebesgue measure,

2 · 3d
Z
m(Eα ) = m(K) ⇒ m(Eα ) ≤ |f (y)|dy.
K⊂Eα cpct α Gα
82

(b) Using Tonelli’s theorem,

Z Z ∞ ∞
√
Z Z
|f ∗ (x)|2 dx = χ|f ∗ (x)|2 >y dydx = m ({x : |f ∗ (x)| > y}) dy.
Rd Rd 0 0
√
Substituting α = y, dy = 2αdα, this equals
Z ∞
2 αm({|f ∗ (x)| > α})dα.
0

(c)


Chapter 6.7, Page 312

Exercise 3: Consider the exterior Lebesgue measure m∗ introduced in Chap-
ter 1. Prove that a set E ⊂ Rd is Carathéodory measurable if and only if
E is Lebesgue measurable in the sense of Chapter 1.
Exercise 4: Let r be a rotation of Rd . Using the fact that the mapping
x 7→ r(x) preserves Lebesgue measure (see Problem 4 in Chapter 2 and
Exercise 26 in Chapter 3), show that it induces a measure-preserving mape
of the sphere S d−1 with its measure dσ.

Solution. Let E ⊂ S d−1 . By definition, σ(E) = dm(Ẽ) where Ẽ is the

union of all radii with endpoints in E. Then if r is a rotation of Rd ,
σ(rE) = dm(rE)˜ by definition. But rE
˜ = rẼ since
˜ ⇔ x = ρθ
x ∈ rE for someρ ≤ 1, θ ∈ rE
⇔ x = ρr(α), ρ ≤ 1, α ∈ E
⇔ x = r(ρα) (since rotations are linear)
⇔ x ∈ r(Ẽ)

Thus, m(rE) = dm(rẼ) = dm(Ẽ) = m(E), so r preserves measures on the

sphere. 

Exercise
R 5: Use2 the polar coordinate formula to prove the following:
(a) Rd e−π|x| dx = 1, when d = 2. Deduce from this that the same
identity holds for all d.
 R ∞ −πr2 d−1
(b) 0
e r dr σ(S d−1 ) = 1, and as a result, σ(S d−1 ) = 2π d/2 /Γ(d/2).
d/2
(c) If B is the
R unit ball,
 vd = m(B) = π /Γ(d/2 + 1), since this quantity
1 d−1
equals 0 r dr σ(S d−1 ).

Solution.
(a) For d = 2, we have by polar coordinates
Z Z Z ∞ Z ∞
2 2 2 2
e−π|x| dx = e−πr rdrdθ = 2π e−πr rdr = −e−πr |∞
0 = 1.
Rd S1 0 0

Note that for general d,

Z Z Z Z Z d
−π|x|d −π(xd +···+xd ) −πxd −πxd −πx2
e dx = e 1 d dx = . . . e 1 ...e d dx1 . . . dxd = e dx
Rd Rd R1
 83

by Tonelli’s Theorem.R Since we have calculated that this equals 1 for

2
d = 2, it follows that R1 e−πx dx = 1, whence the integral over Rd is
1 for all d.
(b) Using integration by parts, for d ≥ 3,
Z ∞ 2 Z ∞ −πr2
−πr 2 d−1 e−πr d−2 ∞ e
e r dr = − r 0
+ (d − 2)rd−3 dr.
0 2π 0 2π
Since σ(S d−1 ) is just the reciprocal of this first integral (which follows
immediately from applying the polar coordinates formula to the result
2π
in part (a)), it follows that σ(S d−1 ) = d−2 σ(S d−3 ). We now prove
the formula σ(S d−1 ) = 2π d/2 /Γ(d/2) by induction. The base cases are
√
σ(S 1 ) = 2π = 2π 2/2 /Γ(2/2) and σ(S 2 ) = 4π = 2π 3/2 /( π/2) =
√
2π 3/2 /Γ(3/2) since Γ(3/2) = 1/2Γ(1/2) = 1/2 π. Now if we let
2π (d−2)/2
ad = 2π d/2 /Γ(d/2), then ad = d−2 π /((d − 2)/2Γ((d − 2)/2) =
2π d−1
a
d−2 d−2 . Since a d and σ(S ) satisfy the same recurrence and initial
conditions, they are equal for all d.
(c) By polar coordinates,
Z 1
d−1 1 2π d−2 π d/2
m(B) = σ(S ) rd−1 dr = 2π d/2 /Γ(d/2) = d = .
0 d 2 2 Γ(d/2) Γ(d/2 + 1)


Exercise 8: The fact that the Lebesgue measure is uniquely characterized by

its translation invariance can be made precise by the following assertion:
If µ is a Borel measure on Rd that is translation-invariant, and is finite
on compact sets, then µ is a multiple of Lebesgue measure m. Prove this
theorem by proceeding as follows.
(a) Suppose Qa denotes a translate of the cube {x : 0 < xj ≤ a, j =
1, . . . , d} of side length a. If we let µ(Q1 ) = c, then µ(Q1/n ) = cn−d
for each integer n.
(b) As a result µ is absolutely continuous with respect to m, and there is
a locally integrable function f such that
Z
µ(E) = f dx.
E

(c) By the differentiation theorem (Corollary 1.7 in Chapter 3) it follows

that f (x) = c a.e., and hence µ = cm.

Solution.
(a) Because Q1 is a disjoint union of nd translates of the cube Q1/n ,
µ(Q1 ) = nd µ(Q1/n ) ⇒ µ(Q1/n ) = n−d µ(Q1 ) = cn−d .
(b) Let E be Borel measurable with m(E) = 0. Then for any  > 0 there
is an open set U with m(U \ E) <  ⇒ m(U ) < . We can write
U as a countable disjoint union of cubes Qj whose side lengths are
of the form 1/n, for example by decomposing U into dyadic rational
cubes as described
P on pp.
P 7-8. Then P µ(Qj ) = cm(Qj ) by part (a),
so µ(U ) = µ(Qj ) = cm(Qj ) = c m(Qj ) = cm(U ) < c. This
can be done for any , so µ(E) = 0. Thus, µ is absolutely continuous
84

wrt m, so there existsR a locally integrable Borel measurable function

f such that µ(E) = E f dm.
(c) Let x be a Lebesgue point of f . Let Qn be a series of dyadic rational
cubes containing x. (Just to be clear, these are “half-open” dyadic
rational cubes, i.e. ones of the form 2mnii ≤ xi < m2in+1
i
for i = 1, . . . , d.)
Then Qn shrinks regularly to x because the ratio of a cube to the
circumscribing ball is constant. For each cube we have
Z
1 1 1
f dµ = µ(Qn ) = cm(Qn ) = c,
m(Qn ) Qn m(Qn ) m(Qn )
so f (x) = c at every Lebesgue point x of f (hence a.e.).

Exercise 9: Let C([a, b]) denote the vector space of continuous functions on
the closed and bounded interval [a, b]. Suppose we are given a Borel measure
µ on this interval, with µ([a, b]) < ∞. Then
Z b
f 7→ `(f ) = f (x)dµ(x)
a
is a linear functional on C([a, b]), with ` positive in the sense that `(f ) ≥ 0
if f ≥ 0.
Prove that, conversely, for any linear functional ` on C([a, b]) that is
positive in the above sense, there is a unique finite Borel measure µ so that
Rb
`(f ) = a f dµ for f ∈ C([a, b]).
Solution. Define the notation f ≺ u to mean 0 ≤ f ≤ 1 and f = 1 on [a, u].
Let
F (u) = ` (f ).
f ≺u
Then F is increasing on [a, b], because for u0 > u, f ≺ u0 ⇒ f ≺ u so F (u0 )
is the infimum of a smaller class of sets. To show F is right continuous, it
suffices to show that for every f ≺ u and  > 0 there exists a u0 > u and
f 0 ≺ u0 with `(f 0 ) < `(f ) + . Let f ≺ u and let C = `(f ). By continuity,
the function (1+ C )f is greater than 1 in some neighborhood u < x < x+δ.
Let f 0 ≺ u + 2δ and f 0 (y) = 0 for y ≥ δ. (Such an f 0 can be constructed, for
example, as piecewise linear, say f 0 = 1 on [a, u+δ/2], f 0 (u+3δ/4) = 0, and
f is linear between δ/2 and 3δ/4.) Then f 0 (x) ≤ (1 + C )f (x) everywhere,
so `(f 0 ) ≤ (1 + C )`(f ) = C + . This proves that F is right continuous.
By Theorem 3.5, there exists a unique Borel measure µ on [a, b] such that
µ((a0 , b0 ]) = F (b0 ) − F (a0 ) for all a ≤ a0 < b0 ≤ b.
Rb
Now we need to show `(f ) = a f dµ for all f ∈ C([a, b]). Let L(f ) =
Rb
a
f dµ. Then it suffices to show `(f ) ≤ L(f ) since this will imply −`(f ) =
`(−f ) ≤ L(−f ) = −L(f ) ⇒ `(f ) ≥ L(f ). Let  > 0. Because continuous
functions can be uniformly approximated by step functions, we may choose
a step function f ≥ f with L(f ) < L(f ) + . Write
X
f = ck χ(ak ,bk ] .
WLOG we may assume P that the intervals (ak , bk ] are disjoint. Choose
a0k > ak and let f0 = ck χ(a0k ,bk ] . Now by the definition of F there
exist continuous gk and hk with gk ≺ bk , hk ≺ a0k , and `(hk ) − F (a0k ) <
 85

`(gk ) − F (bk ) < 2k . WLOG we may also assume f < ck (1 − hk ) on (ak , a0k )
since otherwise we may take
(
0 max(f, ck (1 − hk )) ak < x < a0k
ck (1 − hk ) =
ck a0k ≤ x < bk
and the function h0k defined by these relations will also be continuous, satisfy
h0k ≺ a0k , and have h0k < hk ⇒ `(h0k ) < `(hk ). Now let
X
f˜ = ck (gk − hk ).

Then we have f˜ ≥ f by the above remarks concerning hk . Note also that
X X
`(f˜ ) = ck (`(gk ) − `(hk )) < ck (F (bk ) − F (a0k )) +  = L(f0 ) + .

Since we also have the relations f˜ ≥ f and f ≥ f0 , and both ` and L are
positive,
`(f ) < `(f˜ ) < L(f 0 ) +  < L(f ) +  < L(f ) + 2.


This is true for all , so `(f ) ≤ L(f ). 

Exercise 10: Suppose ν, ν1 , ν2 are signed measures on (X, M) and µ a (pos-
itive) measure on M. Using the symbols ⊥ and  defined in Section 4.2,
prove:
(a) If ν1 ⊥ µ and ν2 ⊥ µ, then ν1 + ν2 ⊥ µ.
(b) If ν1  µ and ν2  µ, then ν1 + ν2  µ.
(c) ν1 ⊥ ν2 implies |ν1 | ⊥ |ν2 |.
(d) ν  |ν|.
(e) If ν ⊥ µ and ν  µ, then ν = 0.
Solution.
(a) Let disjoint A1 and B1 be chosen such that ν1 (E) = ν1 (A1 ∩ E) and
µ(E) = µ(B1 ∩ E) for all measurable E. Similarly, choose A2 and
B2 disjoint with ν2 (E) = ν2 (A2 ∩ E) and µ(E) = µ(B2 ∩ E). Let
A = A1 ∪ A2 and B = B1 ∩ B2 . Note that A and B are disjoint
because A1 ∩ B ⊂ A1 ∩ B1 = ∅ and similarly for A2 . Then for any
measurable E, µ(E) = µ(E ∩ B1 ) = µ(E ∩ B) + µ(E ∩ (B1 \ B2 ).
But µ(B1 \ B2 ) = µ((B1 \ B2 ) ∩ B2 ) = 0, so µ(E) = µ(E ∩ B).
Similarly, ν1 (E) = ν1 (E ∩ A1 ) = ν1 (E ∩ A) − ν1 (E ∩ (A \ A1 )), but
ν1 (A \ A1 ) = ν1 ((A \ A1 ) ∩ A1 ) = 0, so ν1 (E) = ν1 (E ∩ A) and by the
same token, ν2 (E) = ν2 (E ∩ A), so (ν1 + ν2 )(E) = (ν1 + ν2 )(E ∩ A).
Thus, µ and ν1 + ν2 are supported on the disjoint sets A and B.
(b) µ(E) = 0 ⇒ ν1 (E) = ν2 (E) = 0 ⇒ (ν1 + ν2 )(E) = 0.
(c) Choose disjoint A and B such that ν1 (E) = ν1 (E ∩ A) and ν2 (E) =
ν2 (E ∩ B) for all measurable E. Then
X X
|ν1 |(E) = sup |ν1 (Ej )| = sup |ν1 (Ej ∩ A)| = |ν1 |(E ∩ A)|
j j

and similarly |ν2 |(E) = |ν2 |(E ∩ B). Hence |ν1 | and |ν2 | are supported
on the disjoint setsPA and B.
(d) |ν|(E) = 0 ⇒ sup |ν(Ej )| = 0 ⇒ ν(Ej ) = 0 for all subsets Ej ⊂
E ⇒ ν(E) = 0.
86

(e) Let disjoint A and B be chosen with ν(E) = ν(E ∩ A) and µ(E) =
µ(E ∩ B). Then for any measurable E, µ(E ∩ A) = µ((E ∩ A) ∩ B) = 0
because A and B are disjoint. Then ν(E) = ν(E ∩ A) = 0 because
µ(E ∩ A) = 0 and ν  µ.

Exercise 11: Suppose that F is an increasing normalized function on R, and
let F = FA + FC + FJ be the decomposition of F in Exercise 24 of Chapter
3; here FA is absolutely continuous, FC is continuous with FC0 = 0 a.e., and
FJ is a pure jump function. Let µ = µA + µC + µJ with µ, µA , µC , and
µJ the Borel measures associated to F, FA , FC , and FJ respectively. Verify
that:
(i) µA is absolutely continuous with respect to Lebesgue measure and
µA (E) = E F 0 (x)dx for every Lebesgue measurable
R
R set E. R
(ii) RAs a result, if F is absolutely continuous, then f dµ = f dF =
f (x)F 0 (x)dx whenever f and f F 0 are integrable.
(iii) µC + µJ and Lebesgue measure are mutually singular.
Solution.
(i) By definition
X
µA (E) = inf FA (bj ) − FA (aj )
E⊂∪(aj ,bj ]
XZ bj
= inf F 0 (x)dx
E⊂∪(aj ,bj ] aj
Z
≥ inf F 0 (x)dx
E⊂∪(aj ,bj ]
∪(aj ,bj ]
Z
≥ F 0 (x)dx.
E
To prove the reverse inequality, let  > 0 and use the absolute continu-
ity of the integral to find a δ > 0 such that m(E) < δ ⇒ E F 0 (x) < .
R

(In case the assumption that F 0 ∈ L1 is a problem, we can treat

the intersection of E with each interval [n, n + 1) separately.) Now
since E is Lebesgue measurable, there is an open set U ⊃ E such
that m(U \ E) < δ. Let Ū be constructed by writing U as a disjoint
union of open intervals (aRj , bj ) and replacing
R each with (aj , bj ]. Then
m(Ū \E) = m(U \E) and Ū \E F 0 (x)dx = U \E F 0 (x)dx because Ū \E
is U \ E plus countably many points. Thus
Z Z Z Z
F 0 (x)dx = F 0 (x)dx + F 0 (x)dx ≤  + F 0 (x)dx.
Ū Ū \E E E

But Ū is one of the sets over which

R the infimum is taken in the defini-
tion of µAR(E), so µA (E) ≤  + E F 0 (x)dx. This is true for any , so
µA (E) = E F R0 (x)dx. R
(ii) The equation E f dµ = E f F 0 (x)dx follows immediately from (a) in
the case where f is a characteristic function. By the linearity of the
integral, it holds for f a simple function as well. The result for non-
negative f follows from the Monotone Convergence Theorem: Choose
 87

simple fn % f . Then
Z Z Z Z Z Z
f dµ = (lim fn )dµ = lim fn dµ = lim fn (x)F 0 (x)dx = (lim fn )(x)F 0 (x)dx = f (x)F 0 (x)dx.
E E E E E E

Finally, linearity allows us to extend to functions in L1 (µ).

R Note that
f ∈ L (µ) iff f F ∈ L (dx) because both |f |dµ and |fRF 0 |dx are
0
1 1
R

defined as suprema of integrals of simple functions, and |c|dµ =

|gF 0 |dx for g simple so the suprema are taken over the same sets.
R

The condition that f be integrable is, so far as I can tell, superfluous,

unless it means µ-integrable, in which case it’s redundant.
(iii) By Exercise 10a, it is sufficient to prove that µP C and µJ are both
∞
singular wrt Lebesgue measure. Write FJ (x) = k=1 ck χ[xk ,∞) . Let
A = {xk } which is countable and therefore has Lebesgue measure zero,
so that Lebesgue measure is supported on Ac . Now Ac is open, so it
is covered by countably many intervals (aj , bj ]. (We can write Ac as a
countable disjoint union of open intervals, and any open interval is a
countable union of half-closed intervals.) Thus, any subset E ⊂ Ac can
be covered by countably many intervals (aj , bj ] with FJ (aj ) = FJ (bj ),
so µJ (E) = 0. Thus, µJ ⊥ m.
The proof that µC ⊥ m is a bit trickier. (Help from Paul Smith on
this part.) We will use the following lemma, taken from page 35 of
Folland:
Lemma 4. If µF is the Borel measure corresponding to the increasing,
right-continuous function F , then for any µ-measurable set E,
X
µ(E) = inf µ((aj , bj )).
E⊂∪(aj ,bj )

In words, this lemma says that it is equivalent to use coverings of

open intervals instead of half-open intervals. This is nice because it
enables us to use theorems about open covers. The straightforward
but unenlightening proof is in Folland for anyone who cares.
Let A = {x ∈ R : FC0 (x) = 0}c . Then m(A) = 0 by hypothesis, so
Lebesgue measure is supported on Ac . We wish to show that µC (Ac ) =
0, so that µC is supported on A. It is sufficient to show that µC (Ac ∩
[0, 1]) = 0 since Ac = ∪Ac ∩[n, n+1] and replacing FC (x) by FC (x−n)
shifts Ac ∩ [n, n + 1] to Ac ∩ [0, 1]. Thus, let B = Ac ∩ [0, 1]. Let  > 0.
Now for any x ∈ B, FC0 (x) = 0, so ∃hx > 0 such that |FC (y)−FC (x)| <
|y − x| for y ∈ [x − h, x + h]. By the Dreiecksungleichung, this implies
FC (x + h) − FC (x − h) < 2h. Thus, if we let Ix = (x − h, x + h), we
have µ(Ix ) ≤ µ((x−h, x+h]) = FC (x+h)−FC (x−h) < 2h = m(Ix ).
The intervals Ix are an open cover of B, so we can take a countable
subcover In . (Here we use the fact that every subset of R is a Lindelöf
space. The proof-by-jargon of this fact is that every subspace of a
separable metric space is separable, and a metric space is separable
iff it is Lindelöf. The direct proof is that every open subset of R is a
countable union of rational intervals, so for each rational interval we
can take a member of our open cover (if there is any) which contains
that interval, and the resulting countable subcollectionP will still cover
our set.) Next, we shrink the intervals In to make m(In ) < 3.
88

We do this inductively: For a given In , if In ⊂ Ij for some j < n, we

discard In entirely; similarly, if Ij ⊂ In for some j < n then we discard
Ij . Once this is done, In ∩ (∪j<n Ij ) is an open subset of In , hence a
disjoint union of open intervals; but none of these intervals can have
both its endpoints within In because this would imply Ij ⊂ In for some
j < n. Hence if In = (a, b), then In ∩ (∪j<n Ij ) = (a, λ) ∪ (γ, b) for
1
some a ≤ λ < γ ≤ b. By replacing a with max(a, λ − 2n+1 ) and b with
1 0
min(b, γ + 2n+1 ), we form a new interval In ⊂ In with the property
that m(In ∩ (∪j<n Ij )) < 21n . However, ∪nj=1 Ij0 = ∪nj=1 Ij because we
only delete parts of an interval that are covered by other intervals.
We would also like In0 to still have the property that µ(In0 ) < m(In0 ).
Unfortunately, this is only guaranteed as long as In0 contains the central
point xn from the interval In . So far I have not been able to close this
hole in the proof.
Overlooking this problem, we have found an open cover In0 ⊃ B with
the properties that µ(In0 ) < m(In0 ) for each n, and m(In0 ∩(∪j<n Ij0 )) <
1 0 1 3
2n . We may additionally assume that In ⊂ [− 2 , 2 ] for all n since
we
P are only interested in covering [0, 1]. This will then imply that
m(In ) < 3 because if we write In0 = An ∪ Bn where An = In \
(∪j<n Ij0 ) and Bn = In0 ∩ (∪j<n Ij0 ), then the An are disjoint and have
union equal to ∪In0 , and
X X X X 1
m(In0 ) = m(An ) + m(Bn ) = m (∪An ) + m(Bn ) ≤ 2 + =3
2n
where m (∪An ) < 2 because ∪An ⊂ [− 12 , 32 ]. This then implies
X X
µ(B) ≤ µ (∪In0 ) ≤ µ(In0 ) ≤  m(In0 ) < 3.

This is true for any , so µ(B) = 0.



Exercise 14: Suppose (Xj , Mj , µj ), 1 ≤ j ≤ k, is a finite collection of mea-

sure spaces. Show that parallel with the case k = 2 considered in Section 3
one can construct a product measure µ1 ×· · ·×µk on X = X1 ×· · ·×Xk . In
fact, for any set E ⊂ X such that E = E1 × · · · × Ek , with Ej ∈ Mj for all
Qk
j, define µ0 (E) = j=1 µj (Ej ). Verify that µ0 extends to a premeasure on
the algebra A of finite disjoint unions of such sets, and then apply Theorem
1.5.

Solution. First a hand should at least be waved at the fact that A is an

algebra. It is closed under complements because
(E1 × · · · × Ek )c = (E1c × X2 × · · · × Xk )
∪ (E1 × E2c × X3 × · · · × Xk )
∪ · · · ∪ (E1 × E2 · · · × Ek−1 × Ekc ).
This is a “stopping time” argument: we divide the complement into k
sets based on which is the first of the Ek that a point fails to be in. The
intersection of two unions of disjoint measurable rectangles is another union
 89

of disjoint measurable rectangles, so we only need to check unions. This

follows from
(E1 × · · · × Ek ) ∪ (F1 × · · · × Fk ) = (E1 × · · · × Ek )
∪ (F1 \ E1 × F2 × · · · × Fk )
∪ (F1 ∩ E1 × F2 \ E2 × F3 × · · · × Fk )
∪ (F1 ∩ E1 × F2 ∩ E2 × F3 \ E3 × F4 × · · · × Fk )
∪ · · · ∪ (F1 ∩ E1 × · · · × Fk−1 ∩ Ek−1 × Fk \ Ek ).
Finally, to show that the extension from rectangles to a premeasure on
the algebra A generated by them is well-defined, let E1 × · · · × Ek be a
measurable rectangle, and suppose E1 × · · · × Ek = ∪j E1j × · · · × Ekj where
the union is disjoint. This immediately implies
∞
X
χE1 (x1 ) . . . χEk (xk ) = χE j (x1 ) . . . χE j (xj )
1 k
j=1

for all (x1 , . . . , xk ) ∈ X1 ×· · ·×Xk . We can integrate both sides with respect
to x1 , using the monotone convergence theorem to move the integral inside
the sum on the RHS, to obtain
X ∞
µ1 (E1 )χE2 (x2 ) . . . χEk (xk ) = µ1 (E1j )χE j (x2 ) . . . χE j (xk ).
2 k
j=1

We then integrate each side wrt x2 , etc. After doing this k times we obtain
∞
X ∞
X
j j
µ1 (E1 ) . . . µk (Ek ) = µ1 (E1 ) . . . µk (Ek ) ⇒ µ0 (E1 ×· · ·×Ek ) = µ0 (E1j ×· · ·×Ekj )
j=1 j=1

as desired.
Since µ0 is a premeasure on A, it extends to a measure on the σ-algebra
generated by A by Theorem 1.5. 
Exercise 15: The product theory extends to infinitely many factors, under
the requisite assumptions. We consider measure spaces (Xj , Mj , µj ) with
µj (Xj ) = 1 for all but finitely many j. Define a cylinder set E as
{x = (xj ), xj ∈ Ej , Ej ∈ Mj , Ej = Xj for all but finitely many j}.
Q∞
For such a set define µ0 (E) = j=1 µj (Ej ). If A is the algebra generated
by the cylinder sets, µ0 extends to a premeasure on A, and we can apply
Theorem 1.5 again.
Solution. First, note that finite disjoint unions of cylinder sets form an al-
gebra, which is therefore the algebra A. To see this, we can just apply
Exercise 14 because of the condition that finitely many indices inQthe cylin-
der have Ej 6= Xj . For example, to see how unions work, let Q Ej be a
cylinder set (where Ej = Xj for all but finitely many j) and Fj another
cylinder set. Then there are finitely many j for which either Ej or Fj is not
Xj ; we may apply the decomposition from Exercise 14 to these components
while
Q leavingQ the others untouched, and hence obtain a decomposition of
( Ej ) ∪ ( Fj ) into finitely many disjoint cylinder sets. Similar comments
apply to intersections and complements.
90

Q
To verify that µ0 extends to a premeasure on A, let Ej be a cylinder
set, and suppose
∞
Y [∞ Y∞
Ej = Ejk
j=1 k=1 j=1

where the union is disjoint and all but finitely many Ejk are equal to Xj for
any fixed k. The characteristic-function version of this statement is
Y∞ ∞ Y
X ∞
χEj (xj ) = χEjk (xj ).
j=1 k=1 j=1

Integrating both sides with respect to x1 and using the monotone conver-
gence theorem to move the integral inside the sum on the right,
∞
Y ∞
X ∞
Y
µ1 (E1 ) χEj (xj ) = µ1 (E1k ) χEjk (xj ).
j=2 k=1 j=2

Repeating the process ` times, we have

∞
Y ∞
X ∞
Y
µ1 (E1 ) . . . µ` (E` ) χEj (xj ) = µ1 (E1k ) . . . µ` (E`k ) χEjk (xj ).
j=`+1 k=1 j=`+1
Q
As ` → ∞, the LHS approaches µ0 ( Ej ); in fact, it will equal it after
a finite number of steps because all but finitely many Ej equal Xj and
µj (Xj ) = 1 for all but finitely many Xj . For the RHS, we apply monotone
convergence again (this time in `) to see that it approaches
 
∞ Y
X ∞ ∞
X ∞
Y
µj (Ejk ) = µ0  Ejk 
k=1 j=1 k=1 j=1

as desired. Hence µ0 extends to a premeasure on A, and therefore to a

measure on the sigma-algebra generated by A by Theorem 1.5. 
Exercise 16: Consider the d-dimensional torus Td = Rd /Zd . Identify Td
as T1 × · · · × T1 and let µ be the product measure on Td given by µ =
µ1 × · · · × µd , where µj is Lebesgue measure on Xj identified with the circle
T. That is, if we represent each point in Xj uniquely as kj with 0 < xj ≤ 1,
then the measure µj is the induced Lebesgue measure on R restricted to
(0, 1].
(a) Check that the completion µ is Lebesgue measure induced on the cube
Q = {x : 0 < xj ≤ 1, j = 1, . . . , d}.
(b) For each function f on Q let f˜ be its extension to Rd which is periodic,
that is, f˜(x + z) = f˜(x) for every z ∈ Zd . Then f is measurable on Td
if and only if f˜ is measurable on Rd , and f is continuous on Td if and
only if f˜ is continuous on Rd .
d
(c) Suppose f and R g are integrable on T . Show that the integral defining
(f ∗g)(x) = Td f (x−y)g(y)dy is finite for a.e. x, that f ∗g is integrable
over Td , and that f ∗ g = g ∗ f .
(d) For any integrable function f on Td , write
X
f∼ an e2πinẋ
n∈Zd
 91

to mean that an = Td f (x)e−2πinẋ dx. Prove that if g is also integrable,

R

and g ∼ n∈Zd bn e2πinẋ , then

P
X
f ∗g ∼ an bn e2πinẋ .
n∈Zd
2πinẋ
(e) Verify that {e } is an orthonormal basis for L2 (Td ). As a result
kf kL2 (Td ) = n∈Zd |an |2 .
P

(f) Let f be any continuous periodic function on Td . Then f can be uni-

formly approximated by finite linear combinations of the exponentials
{e2πinẋ }n∈Zd .

Solution.
(a) This follows from the translation invariance of µ. To show that the
product of translation invariant measures is translation invariant, let
E = E1 ×· · ·×Ed be a measurable rectangle in Td , and x = (x1 , . . . , xd ) ∈
Td . Then


µ(E+x) = µ (E1 +x1 )×· · ·×(Ed +xd ) = µ1 (E1 +x1 ) . . . µd (Ed +xd ) = µ1 (E1 ) . . . µd (Ed ) = µ(E)
so µ is translation invariant on measurable rectangles. This implies
that the outer measure µ∗ generated by coverings of measurable rect-
angles is also translation invariant. But µ is just the restriction of µ∗
to the sigma-algebra of Carathéodory-measurable sets, so it is trans-
lation invariant as well. This implies that µ is a multiple of Lebesgue
measure; since µ(Td ) = m(Q) = 1, we must have m = µ (modulo the
correspondence between Q and Td ).
(b) This is blindingly obvious, but
f m’ble (resp. cts) ⇔ f −1 (U ) m’ble (resp. open) for open U
⇔ f˜−1 (U ) m’ble (resp. open) in Rd
⇔ f˜ m’ble (resp. cts).

Here we use the fact that f˜−1 (U ) is a lattice consisting of translates

of f −1 (U ) by Zd ; such a set is open or measurable iff f −1 (U ) is.
(c) This is a simple application of Tonelli’s Theorem, exactly analogous
to the case in L1 (Rd ):
Z Z Z
|f ∗ g(x)|dx = f (x − y)g(y)dy dx
Td d d
ZT Z T
≤ |f (x − y)||g(y)|dydx
Td Td
Z Z
= |f (x − y)||g(y)|dxdy
Td Td
= kf kL1 (Td ) kgkL1 (Td )
so f ∗ g is integrable on Td . This in turn implies that it is finite a.e.
Finally, the change of variables u = x − y shows that
Z Z
f ∗ g(x) = f (x − y)g(y)dy = f (u)g(x − u)du = g ∗ f (x).
92

(d) Once again, there is absolutely nothing different from the one-variable
case. Since f ∗g is integrable by our above remarks, and |e−2πin·x | = 1,
f ∗ g(x)e−2πin·x is also integrable, so by Fubini’s theorem
Z Z Z
f ∗ g(x)e−2πin·x dx = e−2πin·x f (x − y)g(y)dydx
Td Td Td
Z Z
= g(y) f (x − y)e−2πin·x dxdy
Td Td
Z Z
−2πin·y
= g(y)e f (x − y)e−2πin·(x−y) dxdy
Td Td
Z
= g(y)e−2πin·y an dy
Td
= an bn .
(e) The orthonormality of this system is evident, since
Z Z d Z
Y d
Y
e−2πin·x e2πim·x dx = e2πi(m−n)·x dx = e2πi(mj −nj )xj dxj = nj
δm j
n
= δm
Td Td j=1 T j=1

where δ is the Kronecker delta function. To show completeness, we

use the fact that an orthonormal system {en } in a Hilbert space is
complete iff hf, en i = 0 for all n ⇒ f = 0. Suppose
Z Z Z Z
−2πin·x −2πin1 x1 −2πin2 xn
0 = hf, e2πin·x
i= f (x)e dx = e e ... e−2πind xd f (x1 , . . . , xd )dxd . . . dx1
Td T T Td
for all n1 , . . . , nd , where the use of Fubini’s theorem is justified by the
integrability of f and the fact that |e−2πink xk | = 1. Let
Z Z Z
F1 (x1 ) = e−2πin2 x2 e−2πin3 x3 . . . e−2πind xd f (x1 , . . . , xd )dxd . . . dx2 .
T T T
Then T e−2πin1 x1 F1 (x1 )dx1 = 0 for all n1 , so F1 (x1 ) = 0 a.e. by the
R

completeness of exponentials in the 1-dimensional case. Now let

Z Z
F2 (x1 , x2 ) = e−2πin3 x3 . . . e−2πind xd f (x1 , . . . , xd )dxd . . . dx1 .
T T
For any fixed R value of x1 , Fn (x1 , x2 ) is a function of x2 with the prop-
erty that T e−2πin2 xn F2 (x1 , x2 )dx2 = F1 (x1 ) = 0 a.e., so we must
have F2 (x1 , x2 ) = 0 for a.e. x2 . Continuing inductively, we see that
f (x1 , . . . , xd ) = 0 for a.e. x1 , . . . , xd .
(f) This problem is begging for the Stone-Weierstrass theorem, but since
we haven’t covered that in class, RI’ll reluctantly do the convolution
stuff. Let gδ (x) = δ1d χ[0,δ]d . Then gδ (x)dx = 1, so
Z Z
|f (x) − f ∗ gδ (x)| = f (x) gδ (y)dy − f (x − y)gδ (y) dy
Z
≤ |f (x) − f (x − y)||gδ (y)|dy.

By the uniform continuity of f , given  > 0 there exists δ0 > 0 such

that |x − y| < δ0 ⇒ |f (x) − f (y)| < . For δ < δ0 ,
Z
|f (x) − f ∗ gδ (x)| ≤ |gδ (y)|dy = .
 93

Thus, f ∗ gδ (x) → f (x) uniformly as δ → 0. However, P if2 an are the

δ
Fourier
P δ 2 coefficients of f and b n those of gδ , then |an | < ∞ and
|bn | < ∞ because f and gδ are in L2 so their Fourier P transforms
are as well. Then by the Cauchy-Schwarz inequality, |an bδn | < ∞.
This implies that f\ ∗ gδ ∈ L1 and 1
P since f ∗ gδ ∈ L , Fourier inversion
holds and we have f ∗ gδ (x) = an bn e2πin·x a.e. (in fact, everywhere,
since both sides are continuous). Now we can choose δ such that
|f − f ∗ gδ | < 2 everywhere. For this δ, since the tails of the convergent
series an bδn go to zero, we can choose some truncation |n|≤N |an bn |
P P

such that |n|>N |an bδn | < 2 . Then for any x,

P

X X
f (x) − an bn e2πin·x ≤ |f (x) − f ∗ gδ (x)| + f ∗ gδ (x) − e2πin·x
|n|≤N |n|≤N

X
= |f (x) − f ∗ gδ (x)| + an bδn e2πin·x
|n>N |
X
≤ |f (x) − f ∗ gδ (x)| + |an bδn |
|n>N |
 
≤ + = .
2 2
Thus, f can be uniformly approximated by trigonometric polynomials.


Exercise 17: By reducing to the case d = 1, show that each “rotation” x 7→

x + α of the torus Td = Rd /Zd is measure preserving, for any α ∈ Rd .

Solution. We first suppose that E ⊂ Td is a measurable rectangle E =

E1 × · · · × Ed where Ek ⊂ T for k = 1, . . . , d. Then
m(τ −1 (E)) = m(E−α) = m((E1 −α1 )×· · ·×(Ed −αd )) = m(E1 ) . . . m(Ed ) = m(E).
Hence τ is measure-preserving on measurable rectangles. But since the
measure of any set is computed in terms of its coverings by measurable
rectangles, τ is measure-preserving on all measurable sets. (We actually use
here the fact that τ −1 is measure-preserving as well; if {Rk } is a covering of
E by measurable rectangles, then {τ −1 (Rk )} is a covering of τ −1 (E) with
the same measure; conversely, if {Rk0 } is a covering of τ −1 (E) by measurable
rectangles, then {τ (Rk0 )} is a covering of E with the same measure.) 

Exercise 18: Suppose τ is a measure-preserving transformation on a measure

space (X, µ) with µ(X) = 1. Recall that a measurable set E is invariant
if τ −1 (E) and E differ by a set of measure zero. A sharper notion is to
require that τ −1 (E) equal E. Prove that if E is any invariant set, there is
a set E 0 so that E 0 = τ −1 (E), and E and E 0 differ by a set of measure zero.

Solution. Let
∞ [
\ ∞
E0 = τ −k (E).
n=1 k=n
94

Then
∞ [
\ ∞
E0 \ E = τ −k (E) \ E
n=1 k=n

and
∞ \
[ ∞
E \ E0 = E \ τ −k (E).
n=1 k=n
−k −k
But E \ τ (E) and τ (E) \ E both have measure zero (this follows from
m(E∆E 0 ) = 0 by an easy induction), and countable unions and intersec-
tions of null sets are null, so m(E∆E 0 ) = 0. Moreover,
∞ [
\ ∞ ∞ [
\ ∞
τ −1 (E 0 ) = τ −1 τ −k (E) = τ −k (E) = E 0
n=1 k=n n=2 k=n

because the sets inside the intersection are nested so we get the same set
whether we start at n = 1 or n = 2. 

Exercise 19: Let τ be a measure-preserving transformation on (X, µ) with

µ(X) = 1. Then τ is ergodic if and only if whenever ν is absolutely con-
tinuous with respect to µ and ν is invariant (that is, ν(τ −1 (E)) = ν(E) for
all measurable sets E), then ν = cµ, with c a constant.

Solution. We use the fact that τ is ergodic iff the only R with f ◦τ =
R functions
f a.e. are constant a.e., as well as the fact that E f dµ = τ −1 (E) f ◦ τ dµ
for measure-preserving maps τ . Let τ be ergodic and let ν  µ be an
invariant measure. By the Radon-Nikodym theorem, dν = hdµ for some
function h ∈ L1 (µ). Then for any measurable E,
Z Z
ν(E) = hdµ = h ◦ τ dµ.
E τ −1 (E)

By the invariance of τ , this equals

Z
ν(τ −1 (E)) = hdµ.
τ −1 (E)

Since this is true for any measurable E, h ◦ τ = h a.e. But since τ is

ergodic, this implies h is constant a.e., so ν = cµ for some constant c.
Conversely, suppose every invariant absolutely continuous measure ν is a
constant times µ. Let f ∈ L1 be any function with the property that
f ◦ τ = f a.e. Define an absolutely continuous measure ν by dν = f dµ.
Then the above calculation (run in reverse) shows that ν is invariant, so
ν = cµ. But this implies that f dµ = dν = cdµ so f is constant a.e. Hence
τ is ergodic. 

Exercise 20: Suppose τ is a measure-preserving transformation on (X, µ). If

µ(τ −n (E) ∩ F ) → µ(E)µ(F )
as n → ∞ for all measurable sets E and F , then (T n f, g) → (f, 1)(1, g)
whenever f, g ∈ L2 (X) with (T f )(x) = f (τ (x)). Thus τ is mixing.
 95

Solution. Suppose µ(τ −n (E) ∩ F ) → µ(E)µ(F ) for measurable E and F .

This means that (T n f, g) → (f, 1)(1, g) if f and g are characteristic func-
tions: say f = χE and g = χF , then
Z Z
(T n f, g) = χE (τ n (x))χF (x)dx = χτ −n (E) χF = m(τ −n (E)∩F ) → µ(E)µ(F ) = (f, g).

By linearity in f and conjugate-linearity in g, this implies (T n f, g) →

(f, 1)(1, g) if f and g are measurable simple functions. Now let f, g ∈ L2
and let fm → f and gm → g be sequences of measurable simple functions.
For each m we have
(T n f, g) = (T n f, g − gm ) + (T n (f − fm ), g) + (T n fm , gm ).
Since T is an isometry, |(T n f, g−gm )| ≤ kT n f kkg−gm k = kf kkg−gm k → 0
as m → ∞. Similarly, (T n (f − fm ), g) → 0 uniformly in n as m → ∞.
n m
Finally, (T n fm , gm ) → (fm , 1)(1, gm ) → (f, 1)(1, g). (To be more precise
about this business of taking limits in two different variables, choose fm
and gm with kfm − f k, kgm − gk < . Then since T is an isometry,
(T n f, g) = (T n fm , gm ) + h(n)
where kh(n)k < 2 for all n. Letting n → ∞, we see that (T n f, g) is even-
tually within 2 of (fm , 1)(1, gm ), which in turn is within C of (f, 1)(1, g)
for some constant C. This is true for all , so (T n f, g) → (f, 1)(g, 1).) 
Exercise 21: Let Td be the torus, and τ : x 7→ x + α the mapping arising
in Exercise 17. Then τ is ergodic if and only if α = (α1 , . . . , αd ) with
α1 , . . . , αd , and 1 are linearly independent over the rationals. To do this
show that:
(a)
m−1 Z
1 X k
f (τ (x)) → f (x)dx
m Td
k=0
as n → ∞, for each x ∈ Td , whenever f is continuous and periodic
and α satisfies the hypothesis.
(b) Prove as a result that in this case τ is uniquely ergodic.
Solution.
(a) Suppose first that α1 , . . . , αd and 1 are dependent over Q, say
p
a1 α1 + · · · + ad αd = ,
q
where αi ∈ Q. Let
1
E = {(x1 , . . . , xd ) : 0 < {q(a1 x1 + · · · + ad xd )} <
},
2
where {z} = z − bzc denotes the fractional part of z. Then m(E) = 12
but E = τ −1 (E), so τ is not ergodic.
On the other hand, suppose α1 , . . . , αd and 1 are independent over Q.
Let f (x) = e2πin·x be any complex exponential. If n = 0, then
m−1 Z
1 X k
f ((τ (x)) = 1 = f (x)dx.
m Td
k=0
96

If n 6= 0, then
m−1 m−1
1 X 1 X e2πin·x 1 − e2πimn·α
f (x)dx = e2πin·x e2πikn·α = .
m m m 1 − e2πin·α
k=0 k=0
2πimn·α
Since |1 − e | ≤ 2, this goes to zero as m → ∞, so
m−1 Z
1 X k
f ((τ (x)) → f (x)dx.
m Td
k=0
Finally, since complex exponentials are uniformly dense in the contin-
uous periodic functions by exercise 16f, the above limit holds for any
continuous periodic function: Let f be a continuous periodic func-
tion and P a finite linear combination of complex exponentials with
|f
R − P | <  everywhere. Choose n sufficiently large that |Am P −
1
Pm−1 k
P dx| <  for all m > n, where Am g(x) = m k=0 g(τ (x)). Then
for m > n,
Z Z Z
Am f (x) − f (x)dx ≤ |Am f (x) − Am P (x)| + Am P (x) − P (x)dx + (P (x) − f (x))dx

<  +  +  = 3.
(b) Unique ergodicity follows by the same logic as the 1-variable case.
Let ν be any invariant measure;Rthen part (a) plus the Mean Ergodic
Theorem shows that Pν (f ) = f dx, where Pν is the projection in
L2 (ν) onto the subspace of invariant functions. This implies that the
image of Pν is just the constant functions. R But we know that the
2
L
R (ν) projection
R of f onto the constants is f dν, so we must have
f dx = f dν for continuous f . Since characteristic functions of
open rectangles can be L2 -approximated by continuous functions, this
implies that m(R) = ν(R) for any open rectangle R. But this implies
that m and ν agree on the Borel sets. Hence m is uniquely ergodic for
this τ .

Exercise 26: There is an L2 version of the maximal ergodic theorem. Sup-
pose τ is a measure-preserving transformation on (X, µ). Here we do not
assume that µ(X) < ∞. Then
m−1
1 X
f ∗ (x) = sup |f (τ k (x))|
m
k=0
satisfies
kf ∗ kL2 (X) ≤ ckf kL2 (X) , whenever f ∈ L2 (X).
The proof is the same as outlined in Problem 6, Chapter 5 for the maximal
function on Rd . With this, extend the pointwise ergodic theorem to the
case where µ(X) = ∞, as follows:
1
Pm−1 k
(a) Show that limm→∞ m k=0 f (τ (x)) converges for a.e. x to P (f )(x)
for every f ∈ L (X), because this holds for a dense subspace of L2 (X).
2

(b) Prove that the conclusion holds for every f ∈ L1 (X), because it holds
for the dense subspace L1 (X) ∩ L2 (X).
 97

Solution.
(a) We use the subspaces S = {f ∈ L2 : f ◦ τ = f } and S1 = {g − T g :
g ∈ L2 } from the proof of the mean ergodic theorem. As shown there,
L2 (X) = S ⊕ S¯1 . Given f ∈ L2 , let  > 0 and write f = f0 + f1 + f2
where f0 ∈ S, f1 + f2 ∈ S¯1 , f1 ∈ S1 , kf2 k < . Since f1 ∈ S1 ,
f1 = g − T g for some g ∈ L2 . Let h = f0 + f1 . Then Am f0 = f0 = P f0
for all m, and
m−1
1 X k 1
Am f1 = T (g − T g) = (g − T m g).
m m
k=0
1
Clearly m g(x) → 0 for all x as m → ∞. Moreover, as shown on page
1 m
301, m T g(x) → 0 for almost all x; one can see this from the fact
that, by the monotone convergence theorem,
Z X ∞ ∞ Z ∞ ∞
1 m 2
X 1 m 2
X 1 m 2
X 1
2
|T (g)(x)| = 2
|T (g)(x)| = 2
kT gk = kgk < ∞.
X m=1 m m=1
m X m=1
m m=1
m2
P 1 m 2
Since m2 |T (g)(x)| is integrable, it is finite almost everywhere,
which means the terms in the series tend to zero for almost all x. The
upshot is that Am f1 (x) → P f1 (x) for a.a. x, so that Am h(x) → P h(x)
a.e. Finally, let
n o
Eα = x ∈ X : lim sup |Am f (x) − P f (x)| > α .
m→∞

Since Am f −P f = Am h−P h+Am (f −h)−P (f −h), Eα ⊂ N ∪Fα ∪Gα ,

where N = {x : Am h(x) 6→ P h(x)},
n αo
Fα = x ∈ X : (f − h)∗ (x) > ,
2
and n αo
Gα = x ∈ X : |P (f − g)(x)| > .
2
Now µ(N ) = 0 as we have already established, and by the L2 maximal
k(f −h)∗ k2 2 2
theorem, µ(Fα ) < (α/2)2 2 < 4cα2 ; similarly, since kP yk ≤ kyk for
kf −hk2 2
y ∈ L2 , µ(Gα ) < (α/2)22 < 4
α2 . Since  is arbitrary, µ(Fα ) = µ(Gα ) =
0 for all α > 0. Thus, µ(Eα ) = 0 for all α > 0, so Am f (x) → P f (x)
a.e.
(b) Let f ∈ L1 . For any  > 0, choose g ∈ L2 ∩ L1 with kf − gk1 < .
Then
Am f (x) − P f (x) = Am g(x) − P g(x) + Am (f − g)(x) − P (f − g)(x).
Let
Eα = {x ∈ X : lim sup |Am f (x) − P f (x)| > 2α} .
Then Eα ⊂ N ∪ Fα ∪ Gα where N = {x : Am g(x) − P g(x) 6→ 0},
Fα = {x ∈ X : lim sup |Am (f − g)(x)| > α} ,
and
Gα = {x ∈ X : lim sup |P (f − g)(x)| > α} .
98

Now µ(N ) = 0 by part (a), and by inequality (24) on page 297,

A A
µ(Fα ) ≤ kf − gk1 < .
α α
Also kP (f − g)k ≤ kf − gk < , so by Chebyshev’s inequality,
1 1
µ(Gα ) ≤ kf − gk1 < .
α α
Since  is arbitrary, this implies µ(Fα ) = µ(Gα ) = µ(Eα ) = 0 for all
α > 0. Hence Am f (x) → P f (x) for a.a. x.


Exercise 27: We saw that if kfn kL2 ≤ 1, then fnn(x) → 0 as n → ∞ for a.e.
x. However, show that the analogue where one replaces the L2 -norm by the
L1 -norm fails, by constructing a sequence {fn }, fn ∈ L1 (X), kfn kL1 ≤ 1,
but with lim sup fnn(x) = ∞ for a.e. x.

Solution. This is yet another example of why Stein & Shakarchi sucks. The
problem doesn’t say anything about conditions on X. In fact, the hint seems
to assume that X = [0, 1]. I will assume that X is σ-finite. I will also assume
that the measure µ has the property that for any measurable set E and any
real number α with 0 ≤ α ≤ µ(E), there is a subset S ⊂ E with µ(S) = α.
(This property holds, for example, in the case of Lebesgue measure, or any
measure which is absolutely continuous with respect to Lebesgue measure.
In fact, we don’t need quite this stringent a requirement–it isn’t necessary
that every subset of X have this nice property, but only that we can find a
nested sequence of subsets of Xn whose measures we can control this way,
where X = ∪Xn and µ(Xn ) < ∞.)
Given these assumptions, let X = ∪Xn where µ(Xn ) < ∞. We construct
a sequence En of measurable subsets with the properties that µ(En ) ≤
1
n log n and that every x ∈ X is in infinitely many En . To do this, we
will construct countably many finite sequences and then string them all
together. The first sequence E2 , . . . , EN1 will have the property that X1 =
1
∪Nj=2 Ej , and µ(Ej ) ≤ j log j . To do this, let E2 be any subset of X1 with
1 1
measure 2 log 2 , unless µ(X1 ) ≤ 2 log 2 , in which case E2 = X1 . Let E3 be
1 1
any subset of X1 \ E2 with measure 3 log 3 , unless µ(X1 \ E2 ) ≤ 3 log 3 , in
which case E3 = X1 \E2 . Let E4 be a subset of X1 \(E2 ∪E3 ) with measure
1 1
4 log 4 , or X1 \ (E2 ∪ E3 ) if this has measure P at most 4 log 4 . This process
1
will terminate in finitely many steps because n log n diverges and µ(X1 )
is finite.
We then construct a second finite sequence of sets EN1 +1 , . . . , EN2 whose
union is X1 ∪ X2 , a third finite sequence whose union is X1 ∪ X2 ∪ X3 , etc.
Let En be the concatenation of all these finite sequences. Then every point
1
in X is in infinitely many En , and µ(En ) ≤ n log n.
Now let fn = n log nχEn . Then kfn k1 = n log nµ(En ) ≤ 1. However,
fn (x)
n = log nχEn (x) and since x is in infinitely many En , lim sup fnn(x) = ∞
for all x. 
 99

Exercise 28: We know by the Borel-Cantelli lemma that P if {En } is a collec-

∞
tion of measurable sets in a measure space (X, µ) and n=1 µ(En ) < ∞,
then E = lim sup{En } has measure zero.
In the opposite direction, if τ is a mixing measure-preserving
P∞ transfor-
mation on X with µ(X) = 1, then whenever n=1 µ(En ) = ∞, there are
integers m = mn so that if En0 = τ −mn (En ), then lim sup(En0 ) = X except
for a set of measure 0.
Solution. Let Fn = Enc . Since
P
µ(En ) = ∞, a theorem on infinite prod-
ucts (e.g. Corollary Q5.6 on page 166 of Conway, Functions of One Complex
Variable) says that µ(Fn ) = 0. Let Fn0 = F1 . Because τ is mixing, there
exists m2 such that
1
µ(τ −m2 (F2 ) ∩ F1 ) < µ(F1 )µ(F2 ) + 2 .
2
n
and let F20 = τ −m (F2 ). Next, choose m3 such that
1
µ(τ −m3 (F3 ) ∩ (F10 ∩ F20 )) < µ(F10 ∩ F20 )µ(F3 ) + 3
2
and
1
µ(τ −m3 (F3 ) ∩ F20 ) < µ(F20 )µ(F3 ) + 3
2
and let F30 = τ −m3 (F3 ). Note that this implies
 
0 0 0 1 1 1 1
µ(F1 ∩F2 ∩F3 ) < µ(F3 ) µ(F1 )µ(F2 ) + 2 + 4 < µ(F1 )µ(F2 )µ(F3 )+ 2 µ(F3 )+ 3
2 2 2 2
and
1
µ(F20 ∩ F30 ) < µ(F2 )µ(F3 ) + .
23
Continuing, we choose mk such that
   
k−1 k−1
\ \ 1
µ τ −mk (Fk ) ∩ Fj0  < µ(Fk )µ  Fj0  + k
2
j=` j=`

for all ` = 1, . . . , k − 1; by induction this is less than

k k k
Y X 1 Y
µ(Fj ) + µ(Fi )
2j i=j+1
j=` j=`+1

where the latter product is taken to be 1 if empty (i.e. when j = k). I

Q∞ prove that this tends to 0 as k → ∞, for any fixed `. As noted before,
will
j=` µ(Fj ) = 0 (any tail of the sum diverges, so any tail
Q∞
of the product
tends to zero) so the first term tends to zero. Similarly, j=i µ(Fj ) = 0 for
any fixed i. Thus, if we split the sum into a sum from ` + 1 to M and a
sum from M to k, the finitely many terms from ` + 1 to M can all be made
arbitrarily small because they each approach zero as k → ∞. On the other
Qk
hand, each term 21j i=j+1 µ(Fi ) is at most 21j , so their sum is less than
P∞ 1
j=M 2j which can be made arbitrarily small by an appropriate choice of
M . This proves that  
∞
\
µ  Fj0  = 0
j=`
100

for all `. Then  

∞ \
[ ∞
µ Fj0  = 0.
`=1 j=`

But the complement of this set is just lim sup Ej0 where Ej0 = τ −mj (Ej ).
Thus, lim sup Ej0 is almost all of X. 

Chapter 6.8, Page 319

Problem 1: Suppose Φ is a C 1 bijection of an open set O in Rd with another
open set O0 in Rd .
(a) If E is a measurable subset of O, then Φ(E) is also measurable.
0
(x)|dx, where Φ0 is the Jacobian of Φ.
R
(b) m(Φ(E)) = E
| det Φ
(c) O0 f (y)dy = O f (Φ(x))| det Φ0 (x)|dx whenever f is integrable on O0 .
R R

Solution.
(a) If K is compact, then Φ(K) is also compact by continuity. Now if A
is Fσ , then A is σ-compact (since every closed set in Rn is a countable
union of compact sets), so Φ(A) is also σ-compact and hence Fσ . Thus,
Φ maps Fσ sets to Fσ sets. Since measurable sets are precisely those
that differ from Fσ sets by a set of measure zero, it suffices to show that
Φ maps sets of measure zero to sets of measure zero. (The following
argument is adopted from Rudin p. 153.) Let E ⊂ O have measure
zero. For each integer n, define
Fn = {x ∈ E : |Φ0 (x)| < n}.

Then for any x ∈ Fn , the definition of derivative implies that |Φ(y)−Φ(x)|

|y−x| <
n for |x − y| < δ for some δ > 0. Define Fn,p ⊂ Fn to be the set of
x for which δ = p1 works. Now m(Fn,p ) = 0 because it’s a subset of
E. I claim that we can cover Fn,p by balls of radius less than p1 with
centers in Fn,p and total measure at most . To do this, we can first
cover Fn,p by an open set of arbitrarily small measure. This open set
can be decomposed into cubes of arbitrarily small diameter, as shown
in chapter 1. If the diameter is sufficiently small (less than a constant
times p1 ), we can cover whichever of these cubes intersect Fn,p with
a ball centered at a point of Fn,p and radius less than p1 ; the other
cubes we discard. The total measure of the resulting balls is at most
a constant times the measure of the open set. (This constant comes
from finding the maximal possible ratio of volumes of the ball covering
one of these small-diameter cubes to the cube, which comes when the
center √
of the ball is at one corner of the cube and the radius of the
ball is 2 times the cube’s diameter.) This proves the claim. Now if
we cover Fn,p by such cubes Bj , centered at xj and with radius rj < p1
then for x ∈ Bj we have |T (x)−T (xj )| ≤ n|x−xj | ≤ nrj . This implies
that
[ X X
T (Fn,p ) ⊂ Bnrj (xj ) ⇒ m(T (Fn,p ) ≤ m(Bnrj (xj )) = nd m(Brj (xj )) < nd .
j
 101

Since  was arbitrary, this shows that m(T (Fn,p )) = 0. Since E =

∪n,p Fn,p , this implies m(Φ(E)) = 0.
(b) I will prove this in the case where E is a closed rectangle contained in
O, and then show that this implies the general case.
Thus, assume E ⊂ R ⊂ U for a compact rectangle R. By absolute
continuity of each of the n components of Φ0 , given any  > 0 there
exists δ > 0 such that |x − y| < δ, for x, y ∈ R, implies |(Φ0 (x) −
Φ0 (y))j | <  for j = 1, . . . , n. Divide R into cubes Qk of diameter less
than δ. Let ak be the center of Qk . Then for x ∈ Qk , the mean value
theorem implies Φ(x)j − Φ(ak )j = Φ0 (c)(x − ak ) for some c on the line
segment connecting ak to k. Then c ∈ Qk so |Φ0 (c) − Φ0 (ak )| < .
Thus, |Φ(x)j − Φ(ak )j − Φ0 (ak )j (x − ak )j | < |x − ak |j . This then
implies kΦ(x) − Φ(ak ) − Φ0 (ak )(x − ak )k < kx − ak k since a vector
that is larger in each component has larger norm. This local statement
translates into the global statement
Φ(ak ) + (1 − )Φ0 (ak )(Qk − ak ) ⊂ Φ(Qk ) ⊂ Φ(ak ) + (1 + )Φ0 (ak )(Qk − ak )
which may be verified by checking it for each x ∈ Qk , since x must
lie in the space between the cubes Φ(ak ) + (1 − )Φ0 (ak )(Qk − ak )
and Φ(ak ) + (1 + )Φ0 (ak )(Qk − ak ). Now Φ is a bijection so Φ(Qk )
are almost disjoint, which means their images are also almost disjoint
since Φ maps sets of measure
P zero to sets of measure zero by part (a).
Thus, m(Φ(∪Qk )) = m(Φ(Qk )) and
X X X
m (Φ(ak ) + (1 − )Φ0 (ak )(Qk − ak )) ≤ m(Φ(Qk )) ≤ m (Φ(ak ) + (1 + )Φ0 (ak )(Qk − ak )) .

Now by Exercise 4 of Chapter 2,

X
m (Φ(ak ) + (1 ± )Φ0 (ak )(Qk − ak )) = (1 ± )| det Φ0 (ak )|m(Qk )

since Φ0 is linear. Thus,

X X X
(1 − ) | det Φ0 (ak )|m(Qk ) ≤ m(Φ(Qk )) ≤ (1 + ) | det Φ0 (ak )|m(Qk ).

| det Φ0 (ak )|m(Qk ) is a Riemann sum for R | det Φ0 (x)|dx, and

P R
But
Riemann integration works because φ0 is continuous on the compact
set R. Thus,
Z Z
(1 − ) | det(Φ0 (x))|dx ≤ m(Φ(R)) ≤ (1 + ) | det(Φ0 (x))|dx
R R

for all , so m(Φ(R)) = R | det(Φ0 (x))|dx.

R

Now let E be any measurable subset of O. Let Un ⊂ O be open

sets containing E with m(Un \ E) < n1 . Then ∩Un = E ∪ N where
m(N ) = 0. Then
Φ(E ∪ N ) = Φ(∩Un ) = ∩Φ(Un ).
(In general it is only true that Φ(∩Un ) ⊂ ∩Φ(Un ), but here we have
equality because Φ is bijective.) Since m(Φ(N )) = 0,
m(Φ(E)) ≤ m(Φ(E ∪ N )) ≤ m(Φ(E)) + m(Φ(N )) ⇒ m(Φ(E ∪ N )) = m(Φ(E)).
102

Now Φ(Un ) are nested sets of finite measure, so

Z Z Z Z
m(∩Φ(Un )) = lim m(Φ(Un )) = lim | det Φ0 | = |Φ0 | = |Φ0 | = |Φ0 |.
Un ∩Un E∪N E
0
R
Thus, we have (finally) that R m(Φ(E)) =R E |Φ |.
(c) We proved in part (b) that O0 g(y)dy = O g(Φ(x))| det Φ0 (x)|dx holds
if g is a characteristic function χE for measurablePE ⊂ O. By linearity,
this extends to a measurable simple function cj χEj . Now for f
nonnegative, take a sequence fn % f of simple functions; then by the
Monotone Convergence Theorem,
Z Z
f (y)dy = lim fn (y)dy
O0 O0
Z
= lim fn (y)dy
0
ZO
= lim fn (Φ(x))| det Φ0 (x)|dx
O
Z
= lim fn (Φ(x))| det Φ0 (x)|dx
O
Z
= f (Φ(x))| det Φ0 (x)|dx.
O

Finally, we can extend to complex integrable f by linearity, since f is

a linear combination of four nonnegative integrable functions.


Problem 2: Show as a consequence of the previous problem: the measure

dµ = dxdy
y2 in the upper half-plane is preserved by any fractional linear
 
a b
transformation z 7→ az+b
cz+d , where belongs to SL2 (R).
c d

Solution. We first note that such a transformation does in fact map the
upper half plane to itself: if z = x + iy, then
az + b a(x + iy) + b ac(x2 + y 2 ) + (ad + bc)x + bd (ad − bc)y
= = 2 2
+i
cz + d c(x + iy) + d (cx + d) + (cy) (cx + d)2 + (cy)2
and since y > 0 and ad − bc = 1 this has positive imaginary part. Now if
0 0
we write the map z 7→ az+b cz+d in terms of its components as (x, y) 7→ (x , y ),
0 0
then using the ugly formulas for x and y from the above expression, we can
compute the even uglier partial derivatives and the Jacobian. However, we
can shortcut that by using the fact that the Jacobian is always the square
norm of the complex derivative. In case this needs proof, suppose z 7→ f (z)
is a complex differentiable function. Then if f 0 (z0 ) = α + βi, the linear
map z 7→ f 0 (z0 )(z − z0 ) can be rewritten as
(x+iy) 7→ (α+βi)((x−x0 )+(y−y0 )) = (α(x−x0 )−β(y−y0 ))+i(β(x−x0 )+α(y−y0 )),
or     
x α −β x − x0
7→
y β α y − y0
 103

which has Jacobian α2 + β 2 = |f 0 (z0 )|2 . Now in our case,

(ad − bc)z 1
f 0 (z) = 2
= .
(cz + d) (cz + d)2
If we let Φ denote the mapping z 7→ z 0 (and equivalently (x, y) 7→ (x0 , y 0 )),
then
Z
1
µ(Φ(E)) = 0 2
dx0 dy 0
Φ(E) (y )
Z
1
= 0 2
| det Φ0 |dxdy
E (y )
((cx + d)2 + (cy)2 )2
Z
= |(c(x + iy) + d)2 |2 dxdy
E y2
((cx + d)2 + (cy)2 )2
Z
= ((cx + d)2 + (cy)2 )2 dxdy
E y2
Z
1
= 2
dxdy
E y
= µ(E).


Problem 3: Let S be a hypersurface in Rd = Rd−1 × R, given by

S = {(x, y) ∈ Rd−1 × R : y = F (x)},
with F a C 1 function defined on an open set Ω ⊂ Rd−1 . For each subset E ⊂
Ω we write Ê for the corresponding subset of S given by Ê = {(x, F (x) :
x ∈ E}. We note that the Borel sets of S can be defined in terms of the
metric on S (which is the restriction of the Euclidean metric on Rd ). Thus
if E is a Borel set in Ω, then Ê is a Borel subset of S.
(a) Let µ be the Borel measure on S given by
Z p
µ(Ê) = 1 + |∇F |2 dx.
E

If B is a ball in Ω, let B̂ = {(x, y) ∈ Rd , d((x, y), B̂) < δ}. Show that
δ

1
µ(B̂) = lim m((B̂)δ ),
δ→0 2δ

where m denotes the d-dimensional Lebesgue measure. This result is

analogous to Theorem 4.4 in Chapter 3.
(b) One may apply (a) to the case when S is the (upper) half of the
unit sphere in Rd , given by y = F (x), F (x) = (1 − |x|2 )1/2 , |x| < 1,
x ∈ Rd−1 . Show that in this case dµ = dσ, the measure on the sphere
arising in the polar coordinate formula in Section 3.2.
(c) The above conclusion allows one to write an explicit formula for dσ in
terms of spherical coordinates. Take, for example, the case d = 3, and
write y = cos θ, x = (x1 , x2 ) = (sin θ cos φ, sin θ sin φ) with 0 ≤ θ < π2 ,
0 ≤ φ < 2π. Then according to (a) and (b) the element of area dσ
equals (1−|x|2 )−1/2 dx. Use the change of variable theorem in Problem
1 to deduce that in this case dσ = sin θdθdφ. This may be generalized
104

to d dimensions, d ≥ 2, to obtain the formulas in Section 2.4 of the

appendix in Book I.

Solution.
(a) We proceed in the steps outlined in Prof. Garnett’s hint:
– Since B is compact and Ωc is closed, the distance between them
is greater than zero, so we can choose δ < d(B, Ωc ). Let Vδ =
{x : d(X, B) < δ} ⊂ Ω. For each x ∈ Vδ define Iδ (x) = {y ∈ R :
(x, y) ∈ B̂ δ } and h(x, δ) = m(Iδ (x)).
– Note that B̂ δ ⊂ Vδ ×R since for (x, y) ∈ Rd−1 ×R, d((x, y), B̂) ≥
d(x, B). By Tonelli’s theorem,
Z Z Z Z
m(B̂ δ ) = χIδ (x) (y) = χIδ (x) (y)dydx = h(x, δ)dx.
Vδ R Vδ
(x,y)∈Vδ ×R

– For x ∈ B, let M = |∇F | evaluated at x. Let ~v be the unit

vector in the direction of ∇F . Then ∇F · ~v = M at x. By the
continuity of ∇F , for any  > 0 we may choose δ0 > 0 such that
|∇F | < M +  and ∇F · ~v > M −  at all pointspwithin δ0 of x.
Suppose δ < δ0 . Then for y ∈ R, if F (x) ≤ y ≤ δ 1 + (M − )2 ,
δ(M −)
consider F (x + t~v ) for 0 ≤ t ≤ 1+(M −)2 . By the construction
of δ, F (x) + (M − )t < F (x + t~v ) < F (x) + (M + )t. Hence,
y(M −)
because 1+(M −)2 <
√ δ(M −) 2 < δ,
1+(M −)

y(M − )2
 
y(M − )
F x+ ~v > F (x) + .
1 + (M − )2 1 + (M − )2
y(M −)
By the intermediate value theorem, there is some t0 < 1+(M −)2
2
y(M −)
at which F (x + t0~v ) = F (x) + 1+(M −)2 . Let x0 = x + t0~
v . Then
(x0 , F (x0 )) ∈ B̂; moreover, the distance squared from (x, y) to
(x0 , F (x0 )) is
2 2
y(M − )2 y2
 
t20 + y − = t2
0 +
1 + (M − )2 1 + (M − )2
2  2
y2

y(M − )
≤ +
1 + (M − )2 1 + (M − )2
y2
= <δ
1 + (M − )2
p
so y ∈ Ix,δ . On the other hand, if y > F (x) + δ 1 + (M + )2 ,
suppose (x, y) ∈ Ix,δ , so there is some x0 with dist((x, y), (x0 , F (x0 ))) <
δ. Clearly this implies |x0 −x| < δ; suppose |x0 −x| = t. Then be-
cause |∇F | < M +  between x and x0 , F (x0 ) < F (x) + (M + )t.
Then the distance squared from (x, y) to (x0 , F (x0 )) is
2
t2 + (y − F (x0 ))2 ≥ t2 + (y − (M + )t) = (1 + (M + )2 )t2 − 2y(M + )t + y 2 .
 105

This is a quadratic polynomial in t; the minimum occurs when

the derivative is zero, i.e. when
y(M + )
2t(1 + (M + )2 ) = 2y(M + ) ⇒ t = .
1 + (M + )2
At this point, the value of the quadratic is
2
y2

2 y(M + ) y(M + ) 2
(1 + (M + ) ) − 2y + y = > δ,
1 + (M + )2 1 + (M + )2 1 + (M + )2
so in fact there is no point in B̂ within δ of (x, y), a contra-
diction. Thus, y ∈ / Ip
x,δ . By symmetry,
p the same results hold
y < F (x), so [−δ p1 + (M − ) , δ 1 + (M − )2 ] ⊂ Ix,δ ⊂
for p 2

[−δ 1 + (M + )2 , δ 1 + (M + )2 ]. Thus,

p h(x, δ) p
1 + (M − )2 ≤ < 1 + (M + )2
2δ
for δ < δ0 . This proves that
h(x, δ) p
lim = 1 + |∇F |2 .
δ→0 2δ
– Because ∇F is continuous on the compact set B, it attains a
maximum M on B. For any x ∈ B, if |y −F (x)| > (M +1)δ then
y∈/ Ix,δ , because any point (x0 , F (x0 )) within δ of (x, y) would
have to have |x0 − x| < δ, and then the bound on ∇F implies
|F (x0 ) − F (x)| < M δ ⇒ |y − F (x0 )| > δ. Thus, h(x,δ)
2δ < M +1
for all δ.
– Since we’re interested in the limit of small δ, we can restrict
our attention toRδ below some R cutoff value, say a, where a <
d(B, Ωc ). Then Vδ h(x, δ) = Va h(x, δ) because h(x, δ) = 0 for
x ∈ Va \ Vδ . This enables us to take all the integrals over the
same region. Moreover, since Va has finite measure, the constant
M + 1 is integrable over Va . So by the dominated convergence
theorem,
Z
1 δ h(x, δ)
lim m(B̂ ) = lim dx
δ→0 2δ δ→0 V 2δ
a
Z
h(x, δ)
= lim dx
2δ
ZVap
= 1 + |∇F |2 dx
B
= µ(B̂).
(The region of integration becomes B in the limit because h(x, δ)
is eventually 0 for x outside B.)
(b) For two points θ, φ ∈ S d−1 , let a(θ, φ) denote the angular distance
between θ and φ, i.e. the angle between the radius vectors to θ and φ.
Similarly, for a point θ ∈ S d−1 and a set E ⊂ S d−1 , define a(θ, E) =
inf φ∈E a(θ, φ). Now given a ball B ⊂ Rd−1 and corresponding ball
106

B̂ ⊂ S d−1 , define Bδ0 = {p ∈ S d−1 : a(p, B̂) < arcsin(δ)}. I claim that
B̂ × [1 − δ, 1 + δ] ⊂ (B̂)δ ⊂ Bδ0 × [1 − δ, 1 + δ].
Here the product is in spherical coordinates, of course. The first inclu-
sion is obvious because if (α, r) ∈ B̂ × [1 − δ, 1 + δ], then (α, 1) ∈ B̂ and
is a distance |1 − r| ≤ δ away. For the second inclusion, let (α, r) be
any point in Rd . f α ∈ / Bδ0 , let (θ, 1) be any point in B̂. The distance
from (θ, 1) to the line through the origin and (α, r) is sin(a(θ, α)) by
elementary trigonometry. By hypothesis this is greater than δ, so the
distance from (θ, 1) to (α, r) is greater than δ. On the other hand,
/ [1 − δ, 1 + δ], then no point in S d−1 is within δ of (α, r). This
if r ∈
proves the second inclusion. Now by the spherical coordinates formulas
derived in section 3,
Z 1+δ
(1 + δ)d − (1 − δ)d
m(B̂ × [1 − δ, 1 + δ]) = σ(B̂) rd−1 dr = σ(B̂),
1−δ d
so
m(B̂ × [1 − δ, 1 + δ]) 1 (1 + δ)d − (1 − δ)d
= σ(B̂).
2δ d 2δ
d d
−(1−δ)
Since (1+δ) 2δ is a difference quotient for the function f (x) = xd
d d 1
at x = 1, it approaches dx x |x=1 = d as δ → 0, so 2δ m(B̂ × [1 − δ, 1 +
δ]) → σ(B̂). Similarly,
1 1 (1 + δ)d − (1 − δ)d
m(Bδ0 × [1 − δ, 1 + δ]) = σ(Bδ0 ).
2δ d 2δ
As δ → 0, this approaches limδ→0 σ(Bδ0 ) (provided the latter exists,
of course). Now since the Bδ0 are nested and ∩Bδ0 = B̂, lim σ(Bδ0 ) =
σ(B̂) = σ(B̂) by the continuity of measures. (It hardly bears pointing
out here that σ(Bδ0 ) are all finite.) By the Sandwich Theorem,
1
µ(B̂ δ ) = lim m(B̂ δ ) = σ(Bδ0 ).
δ→0 2δ

Thus, µ = σ on balls, but since these generate the Borel sets, µ = σ.

(c) By the change of variable theorem,
∂x1 ∂x1
∂θ ∂φ cos θ cos φ − sin θ sin φ
dx = ∂x2 ∂x2 dθdφ = dθdφ = sin θ cos θdθdφ.
∂θ ∂φ
cos θ sin φ sin θ cos φ
Note also that
  !
∂F ∂F −x1 −x2
∇F = , = p ,p
∂x1 ∂x2 1 − x21 − x22 1 − x21 − x22
so
x21 x22 1 1 1
1 + |∇F |2 = 1 + + = = = .
1 − x21 − x22 1 − x21 − x22 1 − x21 − x22 2
1 − sin θ cos 2θ

Then
p 1
dσ = 1 + |∇F |2 dx = sin θ cos θdθdφ = sin θdθdφ.
cos θ
 107

In n dimensions, one may use for spherical coordinates the n angles

θ1 , . . . , θn with
y = cos θ1
x1 = sin θ1 cos θ2
x2 = sin θ1 sin θ2 cos θ3
....
..
xn−1 = sin θ1 . . . sin θn−1 cos θn
xn = sin θ1 . . . sin θn .
∂(x1 ,...,xn )
Then the Jacobian ∂(θ1 ,...,θn ) is
c1 c2 −s1 s2 0 0 ... 0
c1 s2 c3 s1 c2 c3 −s1 s2 s3 0 ... 0
c1 s2 s3 c4 s1 c2 s3 c4 s1 s2 c3 c4 −s1 s2 s3 s4 ... 0
.. .. .. .. .. ..
. . . . . .
c1 s2 . . . sn s1 c2 s3 . . . sn s1 s2 c3 s4 . . . sn s1 s2 s3 c4 s5 . . . sn ... s1 . . . sn−1 cn
where ci = cos θi and si = sin θi . Call this Jn (θ1 , . . . , θn ). Then we
can compute these inductively: a cofactor expansion along the first
row yields
Jn (θ1 , . . . , θn ) = c1 c2 sn−1
1 Jn−1 (θ2 , θ3 , . . . , θn ) + s1 s2 sn−1
2 Jn−1 (θ1 , θ3 , . . . , θn ).
Using as our initial case J2 (θ1 , θ2 ) = c1 s1 as computed above, one has
by an easy induction that
Jn (θ1 , . . . , θn ) = c1 s1n−1 sn−2
2 sn−3
3 . . . sn−1 .
1
Since 1 + |∇F |2 = cos2 θ1 as before, this yields the area element
dσ = sinn−1 θ1 sinn−2 θ2 . . . sin θn−1 dθ1 . . . dθn .

Note: If the point of this exercise was to calculate the area element on
the unit sphere, it seems that a more direct way is to use the change of
variables formula to compute the volume element in spherical coordinates
(r2 sin θdrdθdφ in three dimensions), and since the measure of a set E on
the unit sphere is defined to be the measure of the corresponding conical
segment, we can compute it as
Z Z
σ(E) = 3 r2 sin θdrdθdφ = sin θdθdφ,
[0,1]×E 0 E0
0
where E is the region in the θ − φ plane corresponding to the spherical
region E. This implies dσ = sin θdθdφ, with similar formulas in higher
dimensions.
Problem 6: Consider an automorphism A of Td = Rd /Zd , that is, A is a
linear isomorphism of Rd that preserves the lattice Zd . Note that A can
be written as a d × d matrix whose entries are integers, with det A = ±1.
Define the mapping τ : Td → Td by τ (x) = A(x).
(a) Observe that τ is a measure-preserving isomorphism of Td .
108

(b) Show that τ is ergodic (in fact, mixing) if and only if A has no eigen-
values of the form e2πip/q , where p and q are integers.
(c) Note that τ is never uniquely ergodic. (Hint.)
Solution.
(a) Duly noted.
OK, I guess I’m supposed to prove it. :-) Since A is a linear isomor-
phism, A−1 is as well. Let E ⊂ Td be measurable, and let Ẽ be the
corresponding subset of the unit cube in Rd . Then
µ(A−1 E) = m(A−1 (Ẽ)) = | det A−1 |m(Ẽ) = m(Ẽ) = µ(E)
where µ is the measure on Td induced by the Lebesgue measure m on
Rd .
(b) Suppose that A has an eigenvector of the form e2πip/q ; then AT does
as well, since it has the same characteristic polynomial. Then (AT )q
has 1 as an eigenvector. Since AT − I has all rational entries, there is a
(nonzero) eigenvector in Qd , and hence, by scaling, in Zd . Let n ∈ Zd
with (AT )q n = n. Consider the function f (x) = e2πin·x for x ∈ Td .
T k
Then since n · (Ax) = (AT n) · x, T k f (x) = e2πi((A ) n)·x . The averages
of this function are
m−1
1 X 2πi((AT )k n)
Am f (x) = e .
m
k=0
k+q k
But T f = T f for all k, so Ajq f (x) = Aq f (x) for any integer j.
Since Aq f (x) is not zero (it is a linear combination of exponentials
with distinct periods, since we may assume WLOG that R q is as small
as possible), the averages do not converge a.e. to Td f (x)dx = 0.
Hence τ cannot be ergodic.
On the other hand, suppose A has no eigenvector e2πip/q . Let f (x) =
e2πin·x and g(x) = e2πim·x for any m, n ∈ Zd . Then
Z
T k
hT k f, gi = e2πi((A ) n−m)·x dx.
Td
If m = n = 0 the integrand is 1 and the integral is 1 = hf, gi for all k.
If m and n are not both zero, then (AT )k n − m) is eventually nonzero;
if not, there would be values k1 and k2 at which it were zero, but then
(AT )k1 n = (AT )k2 n so (AT )k1 n is an eigenvector of (AT )k2 −k1 with
eigenvalue 1. (Since A is invertible, this eigenvector is nonzero). This
then implies that A has an eigenvector which is a (k2 − k1 )th root of
unity, a contradiction. Thus, (AT )k n − m is eventually nonzero, so
hT k f, gi is eventually equal to 0 = hf, gi. Hence T is mixing.
As a side note, the fact that (1 is an eigenvalue of Ak ) implies (some kth
root of 1 is an eigenvalue of A) is not completely trivial. The converse
is trivial, of course, but this direction is not if A is not diagonalizable
(at least not for any reason that I’ve found). It follows, however, from
the Jordan canonical form. If Jm (λ) is a Jordan block of size m with
λ on the diagonal, then Jm (λ)k is triangular with λk on the diagonal,
so it has λk as a k-fold eigenvalue. This implies that the algebraic
multiplicity of λk in Ak is the same as the algebraic multiplicity of λ
 109

in A; since A and Ak have the same dimension, Ak can have no other

eigenvalues besides kth powers of eigenvalues of A (since the sum of
the multiplicities of the λki is already equal to the dimension of Ak ).
(c) If ν is the Dirac measure, i.e. ν(E) = 1 if 0 ∈ E and 0 otherwise, then
ν(τ −1 (E)) = ν(E) because the linearity of τ guarantees 0 ∈ τ −1 (E) ⇒
0 ∈ E, and the invertibility of τ guarantees 0 ∈ E ⇒ 0 ∈ τ −1 (E).


Problem 8: Let X = [0, 1), τ (x) = h1/xi, x 6= 0, τ (0) = 0. Here hxi denotes
dx
the fractional part of x. With the measure dµ = log1 2 1+x , we have of course
µ(X) = 1.
Show that τ is a measure-preserving transformation.

Solution. Let (a, b) ⊂ [0, 1). Then

∞ ∞  
1 1 [ [ 1 1
h i ∈ (a, b) ⇔ ∈ (n + a, n + b) ⇔ x ∈ ,
x x n=1 n=1
n+b n+a
so
∞  
X 1 1
µ(τ −1 ((a, b))) = µ ,
n=1
n+b n+a
∞ Z 1/(n+a)
X 1 dx
=
n=1
log 2 1/(n+b) 1 +x
∞ 1
!
1 X 1 + n+a
= log 1
log 2 n=1 1 + n+b
∞
1 Y (n + a + 1)(n + b)
= log .
log 2 n=1
(n + a)(n + b + 1)

But
∞
Y (n + a + 1)(n + b) 1+b
=
n=1
(n + a)(n + b + 1) 1+a
because the product telescopes; all terms cancel except 1 + b on the top
and 1 + a on the bottom. Hence
 
−1 1 1+b
µ(τ ((a, b)) = log = µ((a, b)).
log 2 1+a
Since τ is measure-preserving on intervals and these generate the Borel sets,
it is measure-preserving. 

Note: By following the hint and telescoping a sum rather than a product,
it is possible to prove τ is measure-preserving for all Borel sets directly
rather than proving it for intervals and then passing to all Borel sets. Let
E ⊂ [0, 1) be Borel, let E + k denote the translates of E, and 1/(E + k) =
{1/(x + k) : x ∈ E}. Since
∞ ∞
1 X 1 1 X 1
= − = ,
1+x k+x k+1+x (k + x)(k + 1 + x)
k=1 k=1
110

it follows that
Z
1 1
µ(E) = dx
log 2 E 1 + x
Z X∞
1 1
= dx
log 2 E k=1 (x + k)(x + k + 1)
∞
XZ
1 1
= dx
log 2 (x + k)(x + k + 1)
k=1 E
∞ Z
1 X 1
= dx
log 2 E+k x(x + 1)
k=1

where we have used the monotone convergence theorem to interchange the

sum and the integral. Now if we make the change of variable x = y1 , then
dx dy
it turns out that x(x+1) = 1+y , so this equals
∞ Z ∞  
1 X 1 X 1
dy = µ
log 2 1/(E+k) y+1 E+k
k=1 k=1
∞
!
[ 1
=µ
E+k
k=1
= µ(τ −1 (E)).

Chapter 7.5, Page 380

Exercise 2: Suppose E1 and E2 are two compact subsets of Rd such that
E1 ∩ E2 contains at most one point. Show directly from the definition of
the exterior measure that if 0 < α ≤ d, and E = E1 ∩ E2 , then
m∗α (E) = m∗α (E1 ) + m∗α (E2 ).

Solution. If E∩ E2 = ∅ then d(E1 , E2 ) > 0 because both are compact, so

for δ < d(E1 , E2 ), every δ-cover of E1 ∪ E2 is a disjoint union of a δ-cover
of E1 and a δ-cover of E2 . This implies
Hαδ (E) = Hαδ (E1 ) + Hαδ (E2 )
and taking the limit as δ → 0 yields m∗α (E) = m∗α (E1 ) + m∗α (E2 ).
Now suppose E1 ∩ E2 = {z}. Let F1δ = E1 \ Bδ (z) and F2δ = E2 \ Bδ (z).
For any δ-cover {Fj } of E, let {Ai } be the collection of those Fj which
intersect F1δ , and {Bk } the collection of those that intersect F2δ . Note that
these collections are disjoint because d(F1δ , F2δ ) ≥ δ. Then {Ai } ∪ {Bδ (z)}
is a δ-cover for E1 , and {Bk } ∪ {Bδ (z)} is a δ-cover for E2 . Thus
Hαδ (E1 ) + Hαδ (E2 ) ≤ Hαδ (E) + 2δ α .
Taking limits as δ → 0, we have m∗α (E) ≥ m∗α (E1 ) + m∗α (E2 ). The reverse
inequality always holds, of course, because m∗α is an outer measure. 

Exercise 3: Prove that if f : [0, 1] → R satisfies a Lipschitz condition of

exponent γ > 1, then f is a constant.
 111

Solution. Suppose |f (x)−f (y)| ≤ M |x−y|γ with γ > 1. Let 0 ≤ x < y ≤ 1

and let h = y − x. Then for any integer n,
n−1    
X j+1 j
|f (x) − f (y)| = f x+ h −f x+ h
j=0
n n
n−1    
X j+1 j
≤ f x+ h −f x+ h
j=0
n n
n−1 γ
X h
≤ M
j=0
n
= M hγ n1−γ .
This is true for all n, and the bound approaches 0 as n → ∞, so f (x) =
f (y). 

Exercise 4: Suppose f : [0, 1] → [0, 1] × [0, 1] is surjective and satisfies a

Lipschitz condition
|f (x) − f (y)| ≤ C|x − y|γ .
Prove that γ ≤ 1/2 directly, without using Theorem 2.2.

Solution. Suppose γ > 21 . By constructing a lattice of spacing n1 , we can

find n2 points in [0, 1]2 with the property that any two are at least n1
apart. Call these points yk . For each yk , let xk ∈ [0, 1] be any point in the
 1/γ
|yj −yk |
preimage. Then for any k 6= j, |xj − xk | ≥ C ≥ (Cn)1 1/γ . But for
n sufficiently large, (Cn)1/γ < n2 , so the n2 points xk must all be farther
than n12 from each other. This is manifestly impossible; by the Pigeonhole
Principle, some interval [ nj2 , j+1 2
n2 ] for j = 0, . . . , n − 1 must contain two of
the points. 

Exercise 5: Let f (x) = xk be defined on R, where k is a positive integer,

and let E be a Borel subset of R.
(a) Show that if mα (E) = 0 for some α, then mα (f (E)) = 0.
(b) Prove that dim(E) = dim f (E).

Solution.
(a) Since
∞ ∞
!
[ [
f (E) = f E ∩ [n, n + 1] = f (E ∩ [n, n + 1]),
n=−∞ n=−∞

it is sufficient to show mα (f (E ∩ [n, n + 1])) = 0. But f is Lipschitz

on [n, n + 1] because
|f (x) − f (y)| = |xk − y k | = |x − y||xk−1 + xk−2 y + · · · + xy k−2 + y k−1 |
and the second term is continuous on the compact set [n, n + 1], hence
bounded. By Lemma 2.2, this implies mα (f (E ∩ [n, n + 1])) = 0.
112

(b) Let α > dim E. Then mα (E) = 0 ⇒ mα (f (E)) = 0 so α ≥ dim f (E).

Hence dim f (E) ≤ dim E. To show the reverse, it suffices to show that
mα (f (E)) = 0 implies mα (E) = 0, since then we can apply exactly
the same logic with E and f (E) interchanged. Let
(
(−x)1/k x < 0 and k even
g(x) =
x1/k else.
Note that g(f (x)) = ±x, so E ⊂ g(f (E)) ∪ −g(f (E)). Thus, it will
suffice to prove that g is “σ-Lipschitz”. Since
∞ ∞ ∞   [ ∞  
[ [ [ 1 1 1 1
R= [−n − 1, −n] ∪ [n, n + 1] ∪ − ,− ∪ , ∪ {0},
n=1 n=1 n=1
n n+1 n=1
n+1 n
and g is Lipschitz on each of these compact sets (it is C 1 on all but the
last), the result follows. (If we let {Kn } denote all the sets in the above
decomposition, then g is Lipschitz on Kn , so m∗α (g(f (E)∩Kn )) = 0 by
Lemma 2.2, and by countable additivity m∗α (g(f (E))) = 0. Similarly
m∗α (−g(f (E)) = 0, so m∗α (E) = 0.)

Exercise 6: Let {Ek } be a sequence of Borel sets in Rd . Show that if
dim Ek ≤ α for some α and all k, then
[
dim Ek ≤ α.
k

Solution. Suppose to the contrary that dim ∪Ek > α. Choose α0 with
α < α0 < dim ∪Ek . Then mα0 (∪Ek ) = ∞ because α0 < dim ∪Ek . But
mα0 (Ek ) = 0 for each k because α0 > dim Ek , which implies mα0 (∪Ek ) ≤
P
mα0 (Ek ) = 0 by countable subadditivity. This is a contradiction, so
dim ∪Ek ≤ α.