ON LEIGHTON’S COMPARISON THEOREM

AHMED GHATASHEH AND RUDI WEIKARD

Abstract. We give a simple proof of a fairly flexible comparison theorem

for equations of the type −(p(u0 + su))0 + rp(u0 + su) + qu = 0 on a finite
interval where 1/p, r, s, and q are real and integrable. Flexibility is provided
by two functions which may be chosen freely (within limits) according to the
situation at hand. We illustrate this by presenting some examples and special
cases which include Schrödinger equations with distributional potentials as
well as Jacobi difference equations.

1. Introduction
In 1836 Sturm published his paper [15] containing the celebrated comparison
theorem. For it he studied two equations of the form −(pu0 )0 + qu = 0 to conclude
something about the zeros of the solutions of one equation from the zeros of some
solution of the other equation. In fact, Sturm’s theorem requires p = p̃ > 0, q̃ > q
and the continuity of these coefficients. Then, assuming that −(pũ0 )0 + q̃ũ = 0 and
−(pu)0 + qu = 0, that a and b are two consecutive zeros of ũ, and that ũ and u are
positive in (a, b), one obtains the contradiction
Z b Z b b
(upũ0 − pu0 ũ)0 = upũ0  ≤ 0.

0< (q̃ − q)ũu =
a a a
It follows that, all else being the same, u must have a zero in (a, b). The most
prominent application of this result is in the oscillation theorem which compares
the number of zeros of solutions of the equation −(pu0 )0 + (q − λ)u = 0 for different
values of λ.
Only in 1909 Picone [11] was able to weaken the condition p = p̃. The key was
the identity
 0
ũ p
(p̃ũ0 u − ũpu0 ) = (p̃ − p)ũ02 + (q̃ − q)ũ2 + 2 (ũu0 − ũ0 u)2
u u
now known as Picone’s identity. Assuming 0 < p ≤ p̃ and q < q̃ will yield a similar
result as before by a similar argument.
Another essential improvement is due to Leighton [9] who recognized that it was
enough to require
Z b
[(p − p̃)ũ02 + (q − q̃)ũ2 ] < 0
a
rather than pointwise inequalities.
Recently, perhaps beginning with Savchuk and Shkalikov [12], there has been
an increased interest in Sturm-Liouville equations with distributional coefficients.

Date: February 28, 2017.

Key words and phrases. Comparison theorem; distributional potentials.
Name of TEX file: comp.tex.
1
2 AHMED GHATASHEH AND RUDI WEIKARD

Eckhardt et al. [2] pointed out that all these situations (and more) are covered by
the equation
−(p(u0 + su))0 + rp(u0 + su) + qu = 0 (1.1)
where 1/p, r, s, and q are real and integrable1. Our goal here is to obtain a
generalization of Leighton’s comparison theorem covering two equations of the form
(1.1).
Our main result is Theorem 2.3 which, together with its proof, is contained in
Section 2. In Section 3 we discuss several special cases and examples to illustrate
the use of the main theorem. We also provide an appendix where we gather some
known results for the convenience of the reader.
There are many excellent books concerned with comparison theorems of which
we only mention Swanson [16]. Hinton [6] provides a survey of the subject’s history.
Finally we note that all our integrals are Lebesgue integrals unless indicated
otherwise.

2. The comparison theorem

0 0
Solutions of −(pu ) + qu = 0 are continuously differentiable, if q and 1/p are
continuous. It is then clear that the term ũ/u appearing in Picone’s identity has
finite limits at a and b even if u vanishes there (recall that ũ does, too). The case
where q and 1/p are merely integrable
Rx may be treated by changing the independent
variable according to x 7→ t = a 1/p.
Lemma 2.1. Suppose 1/p, q ∈ L1 ((a, b), R) and that p > 0 almost everywhere. If
there is a non-trivial real-valued function φ which is absolutely continuous on [a, b],
vanishes at a and b, and satisfies
Z b
(pφ02 + qφ2 ) ≤ 0, (2.1)
a
then every real-valued solution of −(pu0 )0 + qu = 0 has a zero in (a, b) unless it is
a constant multiple of φ. The latter case cannot occur when the inequality in (2.1)
is strict.
We emphasize that φ need not be a solution of a differential equation.
Proof. Note that our hypothesis (2.1) implies that pφ02 is integrable since qφ2 is.
Let ψ be a real-valued solution of −(pu0 )0 + qu = 0 which does not vanish in
(a, b). We may assume that ψ > 0 on (a, b).
Define g = pψ 0 φ2 /ψ on (a, b). We claim that g has limit 0 at both a and b which
Rb
implies that g is absolutely continuous on [a, b] and hence that a g 0 = 0. Consider
the behavior of g near a. Our claim is obvious R x when ψ(a) 6= 0, so we assume
ψ(a) = 0. The function k defined by k(x) = a 1/p is absolutely continuous and
strictly increasing. Since k 0 (x) = 0 only on a set of measure 0 it follows from
Lemma B.2 that k −1 is also absolutely continuous (and strictly increasing). Hence,
by Theorem A.1 and Lemma B.1, ψ0 = ψ ◦ k −1 and ψ00 = (pψ 0 ) ◦ k −1 are absolutely
continuous on [0, k(b)]. Since ψ is not the trivial solution, ψ00 (0) > 0 and hence
ψ00 (t) ≥ C for some C > 0 at least when t is in some neighborhood of 0. Therefore
ψ(x) ≥ Ck(x) if x is sufficiently close to a. Next, the Cauchy-Schwarz inequality

1In the case of 1/p we really mean here and below that p is real-valued and 1/p is integrable.
 ON LEIGHTON’S COMPARISON THEOREM 3

Rx Rx
applied to φ(x) = a p−1/2 p1/2 φ0 gives φ(x)2 ≤ k(x) a pφ02 for all x ∈ (a, b).
Therefore, if x is sufficiently close to a,
φ(x)2 1 x 02
Z
0≤ ≤ pφ
ψ(x) C a
which tends to 0 as x tends to a. Since a similar argument works at b the proof of
our claim is complete.
Now g 0 = pφ02 + qφ2 − pψ 2 (φ/ψ)02 (this is a variant of Picone’s identity). Hence
Z b Z b
0≤ pψ 2 (φ/ψ)02 = (pφ02 + qφ2 ) ≤ 0
a a

and so φ/ψ must be constant. Since φ is not trivial, this constant cannot be zero
thus proving the lemma. 

We now extend the previous lemma to cover the general equation (1.1). We
denote the antiderivatives of s and r which vanish at a by S and R, respectively.
Lemma 2.2. Suppose 1/p, q, r, s ∈ L1 ((a, b), R) and that p > 0 almost everywhere.
If there is a non-trivial real-valued function φ which is absolutely continuous on
[a, b], vanishes at a and b, and satisfies
Z b
eS−R (p(φ0 + sφ)2 + qφ2 ) ≤ 0, (2.2)
a

then every real-valued solution of (1.1) has a zero in (a, b) unless it is a constant
multiple of φ. The latter case cannot occur when the inequality in (2.2) is strict.
Proof. Assume ψ solves equation (1.1) and that it is positive on (a, b). Define
p0 = p e−S−R , q0 = q e−S−R , φ0 = φ eS and ψ0 = ψ eS . Then
Z b Z b
eS−R (p(φ0 + sφ)2 + qφ2 ) = (p0 φ02 2
0 + q 0 φ0 )
a a

and
−(p(ψ 0 + sψ))0 + rp(ψ 0 + sψ) + qψ = [−(p0 ψ00 )0 + q0 ψ0 ] eR .
Since ψ0 > 0 on (a, b) the previous lemma shows that ψ0 is a constant multiple of
φ0 and hence ψ a constant multiple of φ. 

Now the question arises of how to find a function φ which satisfies (2.2). The
idea of a comparison theorem is to look for it among the solutions of a related (but
better known) equation with coefficients (p̃, q̃, r̃, s̃). In fact, we will generalize this
idea by multiplying such a solution with a positive absolutely continuous function.
Any such function can be written as eF where F is absolutely continuous and real.
We denote F 0 by f . Thus we set φ = eF ũ where ũ satisfies
−(p̃(ũ0 + s̃ũ))0 + r̃p̃(ũ0 + s̃ũ) + q̃ũ = 0 (2.3)
and ũ(a) = ũ(b) = 0. Condition (2.2) becomes then
Z b
e2F +S−R (p(ũ0 + (f + s)ũ)2 + qũ2 ) ≤ 0. (2.4)
a
4 AHMED GHATASHEH AND RUDI WEIKARD

Now multiply equation (2.3) by eG ũ where G is real and absolutely continuous.

Let g = G0 and note that it is integrable. Then we get after an integration by parts
Z b
0= eG ũ[−(p̃(ũ0 + s̃ũ))0 + r̃p̃(ũ0 + s̃ũ) + q̃ũ]
a
Z b
= eG [(ũ0 + gũ)p̃(ũ0 + s̃ũ) + r̃p̃(ũ0 + s̃ũ)ũ + q̃ũ2 ]
a
Z b
= eG [p̃(ũ0 + s̃ũ)2 + p̃(g + r̃ − s̃)(ũ0 + s̃ũ)ũ + q̃ũ2 ].
a
Subtracting this from (2.4) we obtain the condition
Z b
[A(ũ0 + s̃ũ)2 + B(ũ0 + s̃ũ)ũ + C ũ2 ] ≤ 0 (2.5)
a
where
A = p e2F +S−R −p̃ eG ,
B = 2p(f + s − s̃) e2F +S−R −p̃(g + r̃ − s̃) eG ,
and
C = (q + p(f + s − s̃)2 ) e2F +S−R −q̃ eG .
In (2.5) we tacitly assume the integrability of the integrand. A sufficient condition
for this is the integrability of A/p̃2 , B/p̃, and C.
The following result is now an immediate corollary of Lemma 2.2.
Theorem 2.3 (The Comparison Theorem). Suppose 1/p, 1/p̃, q, q̃, r, r̃, s, s̃ are all
in L1 ((a, b), R), that p and p̃ are positive almost everywhere, and that the differential
equation
−(p̃(u0 + s̃u))0 + r̃p̃(u0 + s̃u) + q̃u = 0
has a non-trivial real solution ũ which vanishes at a and b and satisfies the inequality
(2.5) for some choice of real absolutely continuous functions F and G. Then every
real solution of
−(p(u0 + su))0 + rp(u0 + su) + qu = 0
has a zero in (a, b) unless it is a constant multiple of ũ eF . The latter case cannot
occur when the inequality in (2.5) is strict.

3. Special cases and examples

3.1. The generalized Sturm-Picone theorem. In analogy to S and R we will
also use the antiderivative S̃ of s̃ which vanishes at a.
Theorem 3.1. Suppose 1/p, 1/p̃, q, q̃, s, s̃ ∈ L1 ((a, b), R), that 0 < p ≤ p̃ and q ≤ q̃
almost everywhere, and that
µ = p̃(s − s̃) e−2S
is non-decreasing on [a, b]. If the differential equation
−(p̃(u0 + s̃u))0 + s̃p̃(u0 + s̃u) + q̃u = 0
has a non-trivial real solution ũ with zeros at a and b, then every real solution of
−(p(u0 + su))0 + sp(u0 + su) + qu = 0
 ON LEIGHTON’S COMPARISON THEOREM 5

has a zero in (a, b) unless it is a constant multiple of ũ eS̃−S . The latter case
cannot occur when one of the inequalities p ≤ p̃ or q ≤ q̃ is strict on a set of
positive measure or if µ is strictly increasing on some subinterval of (a, b).
Proof. Choose G = 2F = 2S̃ − 2S and set r = s and r̃ = s̃ in inequality (2.5).
Then A/p̃2 , B/p̃, and C are integrable and we have A = (p − p̃) eG ≤ 0 and
C = (q − q̃) eG ≤ 0. Setting ṽ = ũ eS̃ gives
B(ũ0 + s̃ũ)ũ = µ(ṽ 2 )0 .
Using Theorem 3.36 of Folland [3] we get therefore
Z b Z
B(ũ0 + s̃ũ)ũ = − ṽ 2 dµ ≤ 0.
a [a,b)

Now apply Theorem 2.3. 

We chose r = s and r̃ = s̃ only for simplicity. A similar result holds also in the
general case. More importantly, perhaps, it is not necessary to have µ finite at a
and b. It suffices to assume that µ is non-decreasing on (a, b) and to require
lim µ(x)ũ(x)2 = lim µ(x)ũ(x)2 = 0.
x↓a x↑b

3.2. The generalized Sturm separation theorem. The following is a slight

generalization of Theorem 11.1 in Eckhardt et al. [2] (who have r = s).
Theorem 3.2. Suppose 1/p, q, r, s ∈ L1 ((a, b), R), that p > 0 almost everywhere,
and that the differential equation
−(p(u0 + su))0 + rp(u0 + su) + qu = 0
has real solutions u and ũ. If ũ is non-trivial but has zeros at a and b, then u has
a zero in (a, b) unless it is a constant multiple of ũ.
Proof. Choosing F = 0, G = S − R, p̃ = p, q̃ = q, r̃ = r, and s̃ = s gives
A = B = C = 0 in (2.5). 
3.3. Distributional potentials. Here we consider a Schrödinger equation with a
distributional potential2 to obtain a (slight) generalization of Theorem 2 of Ben
Amara and Shkalikov [13] or Theorem 4.1 of Homa and Hryniv [7]. Note that, for
our approach, incorporating a coefficient p causes only a minor inconvenience at
least when p and 1/p are bounded.
If u is in the Sobolev space W 1,2 ((a, b)) (i.e., u is absolutely continuous on
[a, b] and u0 ∈ L2 ((a, b))), then u00 is a distribution in W −1,2 ((a, b)) defined by
Rb
u00 (φ) = − a u0 φ0 . Moreover, if u ∈ W 1,2 ((a, b)) and if v ∈ W −1,2 ((a, b)) has
Rb
antiderivative V ∈ L2 ((a, b)), then we may define (vu)(φ) = − a V (uφ)0 which
shows that vu is also in W −1,2 ((a, b)). We may therefore pose the differential
equation
−u00 + vu = 0.
Thus u ∈ W 1,2 ((a, b)) is a solution of this equation, if, for all test functions φ,
Z b Z b
00 0 0 0
0 = (−u + vu)(φ) = ((u − V u)φ − V u φ) = (u0 − V u + W )φ0
a a

2Appendix C gathers some basic facts about distributions.

6 AHMED GHATASHEH AND RUDI WEIKARD

where W is an antiderivative of the integrable function V u0 and hence absolutely

continuous. By Du Bois-Reymond’s lemma C.1 u0 − V u + W is constant which
implies that u0 − V u is absolutely continuous and hence (u0 − V u)0 = −W 0 =
−V u0 = −V (u0 − V u) − V 2 u almost everywhere. It follows that u satisfies equation
(1.1) with p = 1, q = −V 2 and r = s = −V . Conversely, since s = −V ∈ L2 ((a, b)),
a solution of (1.1) is necessarily in W 1,2 ((a, b)) and solves −u00 + vu = 0 in the
sense of distributions.
Theorem 3.3. Suppose v and ṽ are real distributions in W −1,2 ((a, b)), that ṽ − v
is a non-negative measure, and that the differential equation −u00 + ṽu = 0 has
a non-trivial real solution ũ with zeros at a and b. Then every real solution of
−u00 + vu = 0 has a zero in (a, b) unless it is a constant multiple of ũ.
Proof. Since ṽ − v is a non-negative measure we have that µ = Ṽ − V is non-
decreasing. With F = G = 0 we have A = Bs̃ + C = 0 in inequality (2.5) which
then becomes
Z b Z b Z
2(Ṽ − V )ũ0 ũ = µ(ũ2 )0 = − ũ2 dµ ≤ 0
a a [a,b)

using again Theorem 3.36 of Folland [3]. 

3.4. Difference equations. A comparison theorem for the Jacobi difference equa-
tion is known at least since the work of Fort [4] in 1948. However, it may be viewed
as a special case of Theorem 2.3 as we will show now.
Let α be a sequence of positive numbers defined on N0 and β a sequence of real
numbers defined on N. We consider the difference equation
αn−1 un−1 + βn un + αn un+1 = 0, N0 + 1 ≤ n ≤ N1 − 1 (3.1)
3
on a bounded interval [N0 , N1 ] of N0 for which N1 − N0 ≥ 2. One might want to
write equation (3.1) in terms of forward differences un+1 − un . It then reads
−αn (un+1 − un ) + αn−1 (un − un−1 ) + vn un = 0, N0 + 1 ≤ n ≤ N1 − 1
where vn = −βn − αn − αn−1 .
A solution u : [N0 , N1 ] → R of (3.1) may change sign without ever being zero.
We will therefore be interested in sign changes rather than zeros of solutions. To
be precise we will make the following definition.
Definition 3.4. The sequence u : [N0 , N1 ] → R changes sign on [N0 , N1 ] if there
are n, m ∈ [N0 , N1 ] such that un um < 0.
If u is a real non-trivial solution of (3.1) and un = 0 for some n ∈ [N0 +1, N1 −1],
then un−1 and un+1 must have different signs. In other words, if u has a zero in
[N0 + 1, N1 − 1], then it changes sign on [N0 , N1 ].
Theorem 3.5. Suppose α and α̃ are positive sequences defined on a bounded in-
terval [N0 , N1 − 1] and that v and ṽ are real sequences on [N0 + 1, N1 − 1]. If the
difference equation
−α̃n (un+1 − un ) + α̃n−1 (un − un−1 ) + ṽn un = 0, N0 + 1 ≤ n ≤ N1 − 1 (3.2)

3All closed intervals in this section are to be viewed as subsets of N while half open intervals
0
are considered subsets of R.
 ON LEIGHTON’S COMPARISON THEOREM 7

has a non-trivial solution ũ : [N0 , N1 ] → R such that ũN0 = 0, ũN1 −1 ũN1 ≤ 0, and
1 −1
NX
(αn−1 − α̃n−1 )(ũn − ũn−1 )2 + (vn − ṽn )ũ2n
 

n=N0 +1

+ (αN1 −1 − α̃N1 −1 )(ũ2N1 −1 − ũN1 −1 ũN1 ) ≤ 0 (3.3)

then every real solution of (3.1) changes sign on [N0 , N1 ] unless it is a constant
multiple of ũ. The latter case cannot occur when the inequality in (3.3) is strict.
Proof. We define each of p̃, q̃, and s̃ = r̃ P
on the real interval [N0 , N1 ) to be piecewise
n
constant, specifically p̃ = α̃n and s̃ = − k=N0 +1 ṽk /α̃n on [n, n + 1) (in particular,
2
s̃ = 0 on [N0 , N0 + 1)) and q̃ = −p̃s̃ . Moreover, we define ũ to be continuous and
piecewise linear assuming the given values at the points of [N0 , N1 ]. Thus, on [n, n+
1) we have ũ(x) = ũn + (ũn+1 − ũn )(x − n). Since n 7→ ũn satisfies the difference
equation (3.2) the function x 7→ ũ(x) satisfies −(p̃(ũ0 + s̃ũ))0 + s̃p̃(ũ0 + s̃ũ) + q̃ũ = 0.
Analogously, we define p, q and s = r with the aid of the coefficients αn and vn . Now
set a = N0 and b to be the point where the straight line segment joining the points
(N1 − 1, ũN1 −1 ) and (N1 , ũN1 ) crosses the abscissa. Then, choosing F = G = 0,
the left-hand side of inequality (2.5) equals the left-hand side of inequality (3.3).
Since the solutions of the differential equation (1.1) and the difference equation
(3.1) are in one-to-one correspondence we may apply Theorem 2.3 to obtain our
conclusion. 
3.5. Leighton’s example. Leighton [9] discusses the following example to illus-
trate the use of his integral condition as compared to Picone’s pointwise condition.
Here we show that a careful choice of F and G gives another improvement.
The function ũ(x) = sin(x) solves the equation −u00 − u = 0. With its help we
want to draw conclusions about the zeros of the solutions of −u00 + qu = 0 when
q = k − 1 − x. Using this in (2.5) with the choice 2F = G we get A = B = 0
and
R π C(x) = (k − x + g(x)2 /4) eG(x) . With Leighton’s choice G = 0 we obtain
2
0
C(x) sin(x) dx = π(2k − π)/4. If k > π/2 this is positive and does not allow
Rπ
a conclusion. With G = 0.6x and k ≤ 1.672 we obtain that 0 C(x) sin(x)2 dx is
negative. In this case every solution of −u00 + qu = 0 has a zero in (0, π). Note that
for k ≥ 1.676 one can find solutions without zeros in (0, π).

Appendix A. Existence and uniqueness

The following existence and uniqueness theorem is an important ingredient of
our results.
Theorem A.1. Suppose 1/p, q, r and s are real-valued and integrable on a bounded
interval (a, b), x0 ∈ [a, b], and A, B ∈ R. Then there is a unique real-valued solution
u of the differential equation
−(p(u0 + su))0 + rp(u0 + su) + qu = 0
such that u and p(u0 + su) are absolutely continuous on [a, b] and satisfy the initial
conditions u(x0 ) = A and [p(u0 + su)](x0 ) = B.
Actually p(u0 + su) is, in general, only almost everywhere equal to an absolutely
continuous function. Similarly the equality in the differential equation may only
hold almost everywhere.
8 AHMED GHATASHEH AND RUDI WEIKARD

The theorem follows from a standard iteration scheme since the equation is
equivalent to the system U 0 = M U where u is the first component of U , and
 
−s 1/p
M=
q r
is integrable.

Appendix B. Absolutely continuous functions

Since we need one or two facts about absolutely continuous functions whose
proofs do not seem readily available we provide this appendix.
While sums and products of absolutely continuous functions are again abso-
lutely continuous, the same cannot necessarily be said for compositions. Instead,
the composition of two absolutely continuous functions is absolutely continuous if
and only if it is of bounded variation (see Natanson [10], Theorem IX.3.5). The fol-
lowing special cases suffice for our purposes and may be proved in a straightforward
manner.
Lemma B.1. Suppose f : [α, β] → [a, b] and g : [a, b] → [A, B] are absolutely
continuous, that f is strictly increasing, and that h : [A, B] → R is Lipschitz. Then
g ◦ f : [α, β] → [A, B] and h ◦ g : [a, b] → R are also absolutely continuous.
Related to this, but apparently less well known, is the question of the absolute
continuity of the inverse of an absolutely continuous function. Spataru [14] gave an
example of a strictly increasing absolutely continuous function whose inverse is not
absolutely continuous.
Lemma B.2. If f : [a, b] → R is strictly increasing and continuous the following
statements hold.
(1) f is absolutely continuous on [a, b] if and only if the Lebesgue measure of
f ({x ∈ [a, b] : f 0 (x) = ∞}) equals 0.
(2) f −1 is absolutely continuous on [f (a), f (b)] if and only if the Lebesgue mea-
sure of {x ∈ [a, b] : f 0 (x) = 0} equals 0.
These results are stated as an exercise in Natanson [10]. Their proofs require the
notion of a derived number and an additional result (Lemma B.3 below), which in
turn, relies on Vitali’s covering theorem. We denote Lebesgue measure by m and
the corresponding outer measure by m∗ .
Let the function f : [a, b] → R and the point x0 ∈ [a, b] be given. If n 7→ xn ∈
[a, b] \ {x0 } converges to x0 , the sequence n 7→ (f (xn ) − f (x0 ))/(xn − x0 ) has at
least one limit point in [−∞, ∞]. Any such limit point is called a derived number
for f at x0 . Clearly, f is differentiable at x0 (allowing ±∞ as derivatives), if all
derived numbers coincide.
The following is Lemma 7.1 in Bruckner, Bruckner, and Thomson [1].
Lemma B.3. Let f : [a, b] → R be strictly increasing and let E be a subset of
[a, b]. If at each point x ∈ E there exists a derived number not exceeding p, then
m∗ (f (E)) ≤ p m∗ (E).
Proof of Lemma B.2. We will first prove the only if direction of (1), and then the
if direction. Finally we will show that (1) and (2) are equivalent. First note that f
is of bounded variation and thus has a finite derivative almost everywhere. Since
f is increasing all derived numbers are non-negative. Let A be the set of those x
 ON LEIGHTON’S COMPARISON THEOREM 9

where f 0 (x) exists and is finite, B the set where f 0 (x) = ∞, and C the set where
no derivative exists (not even an infinite one). These sets are pairwise disjoint and
their union is [a, b]. We know that m(B) = m(C) = 0.
Now assume that f is absolutely continuous. Then it maps sets of measure 0 to
sets of measure 0 and, in particular, m(f (B)) = 0, completing our first step.
For the second step m(f (B)) = 0 is the assumption and our main objective is to
show that m(f (C)) = 0 but first we need to investigate whether images of measur-
able sets are measurable. To this end let A = {E : f (E) is Lebesgue measurable}.
It is easy to see that A is a σ-algebra in [a, b] which contains all relatively open
subintervals of [a, b] and hence all its Borel sets. In particular, the image of any
Borel set is Lebesgue measurable. Thus E 7→ µ(E) = m(f (E)) is a measure de-
fined on the Borel sets of [a, b]. Since µ([a, x)) = f (x) − f (a) we find that µ is the
Lebesgue-Stieltjes measure generated by f (or rather the restriction of this to the
Borel sets). Recall that f is an absolutely continuous function if and only if µ is
absolutely continuous with respect to Lebesgue measure. By the Radon-Nikodym
theorem µ = µa + µs where µa is absolutely continuous with respect to m while µs
and m are mutually singular. Thus we need to show that µs = 0.
Since µs and m are mutually singular, there is a set S ⊂ [a, b] such that m(S) =
0 = µs ([a, b] \ S). Since µ(B) = 0 and m(C) = 0 we may assume that B ⊂ S c
and C ⊂ S. For n ∈ N let Sn be the set of those S∞x ∈ S for which there is a
derived number for f smaller than n. Then S = n=1 Sn . But, by Lemma B.3,
µs (Sn ) = µ(Sn ) ≤ n m(Sn ) = 0 which implies µs = 0. We have now proved the
first statement.
To prove that the second statement is equivalent to the first let g = f −1 . Then
note that g is strictly increasing and continuous and that g({t ∈ [f (a), f (b)] : g 0 (t) =
∞}) = {x ∈ [a, b] : f 0 (t) = 0}. 

Appendix C. Distributions
Distributions are linear functionals on a set of test functions4. Specifically, in our
context, test functions are complex-valued, compactly supported functions defined
on (a, b) which have derivatives of all orders. The space of test functions is denoted
by D((a, b)). A linear functional u on D((a, b)) is a distribution, if for every compact
set K ⊂ (a, b) there are constants C > 0 and k ∈ N0 such that
k
X
|u(φ)| ≤ C sup{|φ(j) (x)| : x ∈ K} (C.1)
j=0

whenever the test function φ has its support in K. The set of all distributions on
(a, b) is denoted by D0 ((a, b)). It becomes a linear space upon defining αu1 + βu2
by (αu1 + βu2 )(φ) = αu1 (φ) + βu2 (φ) whenever u1 , u2 ∈ D0 ((a, b)) and α, β ∈ C.
If u is a distribution then so is φ 7→ −u(φ0 ). This distribution is called the
derivative of u and is denoted by u0 . Distributions also have antiderivatives. To
Rb Rx Rb
see this fix ψ ∈ D((a, b)) with a ψ = 1 so that ϕ(x) = a (φ − ψ a φ) defines a
test function in D((a, b)) whenever φ is one. Now, if u is a distribution, define the
linear functional v : φ 7→ −u(ϕ). It is easy to check that v is a distribution. Also,
Rb
since a φ0 = 0, we find that v 0 (φ) = −v(φ0 ) = u(φ), i.e., v is an antiderivative of

4See Gelfand and Shilov [5] or Hörmander [8] for details going beyond the very basics discussed
here.
10 AHMED GHATASHEH AND RUDI WEIKARD

u. Two antiderivatives of a distribution differ by only a constant as the following

lemma shows.

Lemma C.1 (Du Bois-Reymond). Suppose the derivative of the distribution v is

zero. Then v is the constant distribution, i.e., there is a constant C such that
Rb
v(φ) = C a φ for all φ ∈ D((a, b)).

A distribution u is called real if u(φ) is real whenever φ assumes only real values.
It is called non-negative if u(φ) ≥ 0 whenever φ ≥ 0.
We conclude by giving some pertinent examples of distributions. R A complex
measure µ on [a, b] may be identified with the distribution φ 7→ [a,b] φ dµ. In
particular, if µ is absolutely continuous with respect to Lebesgue measure m and
if f ∈ L1 ((a, b)) is the corresponding Radon-Nikodym derivative, we have the
Rb R
distribution φ 7→ a f φ dm. The antiderivative of φ 7→ [a,b] φ dµ is given by
R
φ 7→ [a,b] F φ dm where F (x) = µ([a, x)) is a function of bounded variation.
Rb
The class W −1,2 ((a, b)) consists of the distributions defined by φ 7→ − a F φ0
where F ∈ L2 ((a, b)). Note that the complex measures are a subset of W −1,2 ((a, b)).

References
[1] Andrew M. Bruckner, Judith B. Bruckner, and Brian S. Thomson. Real Analysis. Classical-
RealAnalysis.com, 2008. Second edition.
[2] Jonathan Eckhardt, Fritz Gesztesy, Roger Nichols, and Gerald Teschl. Weyl-Titchmarsh the-
ory for Sturm-Liouville operators with distributional potentials. Opuscula Math., 33(3):467–
563, 2013.
[3] Gerald B. Folland. Real analysis. Pure and Applied Mathematics (New York). John Wiley
& Sons, Inc., New York, second edition, 1999. Modern techniques and their applications, A
Wiley-Interscience Publication.
[4] Tomlinson Fort. Finite Differences and Difference Equations in the Real Domain. Oxford,
at the Clarendon Press, 1948.
[5] I. M. Gel0 fand and G. E. Shilov. Generalized functions. Vol. 1. Academic Press [Harcourt
Brace Jovanovich Publishers], New York, 1964 [1977]. Properties and operations, Translated
from the Russian by Eugene Saletan.
[6] Don Hinton. Sturm’s 1836 oscillation results evolution of the theory. In Sturm-Liouville the-
ory. Past and present. Selected survey articles based on lectures presented at a colloquium and
workshop in Geneva, Italy, September 15–19, 2003 to commemorate the 200th anniversary
of the birth of Charles François Sturm, pages 1–27. Basel: Birkhäuser, 2005.
[7] Monika Homa and Rostyslav Hryniv. Comparison and oscillation theorems for singular Sturm-
Liouville operators. Opuscula Math., 34(1):97–113, 2014.
[8] Lars Hörmander. The analysis of linear partial differential operators. I. Classics in Mathe-
matics. Springer-Verlag, Berlin, 2003. Distribution theory and Fourier analysis, Reprint of
the second (1990) edition [Springer, Berlin; MR1065993 (91m:35001a)].
[9] Walter Leighton. Comparison theorems for linear differential equations of second order. Proc.
Amer. Math. Soc., 13:603–610, 1962.
[10] I. P. Natanson. Theory of functions of a real variable. Frederick Ungar Publishing Co., New
York, 1955. Translated by Leo F. Boron with the collaboration of Edwin Hewitt.
[11] Mauro Picone. Sui valori eccezionali di un parametro da cui dipende un’equazione differenziale
lineare ordinaria del second’ordine. Annali della Scuola Normale Superiore di Pisa - Classe
di Scienze, 11:1–144, 1910.
[12] A. M. Savchuk and A. A. Shkalikov. Sturm-Liouville operators with singular potentials. Math-
ematical Notes, 66(6):741–753, 1999. Translated from Mat. Zametki, Vol. 66, pp. 897–912
(1999).
[13] A. A. Shkalikov and Zh. Ben Amara. Oscillation theorems for Sturm-Liouville problems with
distribution potentials. Vestnik Moskov. Univ. Ser. I Mat. Mekh., (3):43–49, 2009.
 ON LEIGHTON’S COMPARISON THEOREM 11

[14] Silvia Spătaru. An absolutely continuous function whose inverse function is not absolutely
continuous. Note Mat., 23(1):47–49, 2004/05.
[15] Sturm. Mmoire sur les quations diffrentielles linaires du second ordre. Journal de Mathma-
tiques Pures et Appliques, 1:106–186, 1836.
[16] C.A. Swanson. Comparison and oscillation theory of linear differential equations. Academic
Press, 1968.

A.G.: Department of Mathematics, University of Alabama at Birmingham, Birming-

ham, AL 35226-1170, USA
E-mail address: ghatash@uab.edu

R.W.: Department of Mathematics, University of Alabama at Birmingham, Birming-

ham, AL 35226-1170, USA
E-mail address: rudi@math.uab.edu