SOME REMARKS ON POINTS OF LEBESGUE DENSITY

AND DENSITY-DEGREE FUNCTIONS

SILVANO DELLADIO

Abstract. Some properties of m-density points and density-degree functions are stud-
ied. Moreover the following main results are provided:
• Let λ be a continuous differential form of degree h in Rn (with h ≥ 0) having the
following property: There exists a continuous differential form ∆ of degree h + 1 in
Rn such that
Z Z
∆∧ω = λ ∧ dω,
Rn Rn
for every Cc∞ differential form ω of degree n − h − 1 in Rn . Moreover let µ be a C 1
differential form of degree h + 1 in Rn and set E := {y ∈ Rn | ∆(y) = µ(y)}. Then
dµ(x) = 0 whenever x is a (n + 1)-density point of E.
• Let f : Rn → R be a measurable function such that f (x) ∈ {0} ∪ [n, +∞] for
a.e. x ∈ Rn . Then there exists a countable family {Fk }∞k=1 of closed subsets of R
n
∞
such that the corresponding sequence of density-degree functions {dFk }k=1 converges
almost everywhere to f .

1. Introduction

Let m ∈ [n, +∞) and E ⊂ Rn . Then x ∈ Rn is said to be an m-density point of E if

Ln (B(x, r) \ E) = o(rm ) as r → 0+ , where Ln denotes the Lebesgue outer measure in
Rn and B(x, r) is the open ball of radius r centered at x. The set of m-density points
of E is denoted by E (m) (cf. [2, 3]). For x ∈ E (n) , the supremum of the numbers m
such that x ∈ E (m) is called the density-degree of E at x and is denoted by dE (x),
while dE (x) is assumed to be zero when x 6∈ E (n) (cf. [7]). Thanks to these definitions,
points of Lebesgue density cease to be indistinguishable from each other and are instead
characterised by their own density-degree.

There are reasons for considering points with a high density-degree, albeit finite, as in-
terior points. For example, the following fact holds: If x ∈ {grad f = F }(n+1) , where
f ∈ C 1 (Rn ) and F ∈ C 1 (Rn , Rn ), then the Jacobian matrix DF (x) is symmetric. We
thus discover that the geometry of {grad f = F } is characterised by a very low density
at points where DF is non-symmetric (cf. [2, Theorem 2.1]; subsequent extensions to the
2020 Mathematics Subject Classification. Primary 28A75, 28A05, 31C40 .
Key words and phrases. Points of Lebesgue density, Density-degree function, Superdensity.

1
2 SILVANO DELLADIO

context of PDE can be found e.g. in [10, Theorem 3.6] and [11, Theorem 3.2]).
A similar application can be given in the context of the Frobenius theorem on distribu-
tions. More precisely let M and D be, respectively, a n-dimensional C 1 submanifold of
Rn+k and a C 1 distribution of rank n on Rn+k . Then D is involutive at each (n+1)-density
point of the tangency set of M with respect to D. Hence the tangency must be low in
density at all points where the distribution D is non-involutive (cf. [8, Theorem 1.1]).

Another fact worth mentioning in this introduction is the following: Except for a subset
of null measure, the points of a set of locally finite perimeter are m-density points, with
1
m := n + 1 + n−1 (cf. [2, Lemma 4.1]; for further applications in this context, see [5, 6]).
Using m-density one can also define the notion of m-approximate continuity, which for
m = n reduces to the well known approximate continuity (cf. [13, Section 2.9.12] and [12,
Section 1.7.2]). Properties of m-approximate continuity holds for Sobolev functions and
for functions of bounded variation (cf. [9]).

Finally, let us briefly describe the two main results of this note. The first one, proved in
Section 3 below, is the following generalization of [2, Theorem 2.1].
Theorem. Let λ be a continuous differential form of degree h in Rn (with h ≥ 0) having
the following property: There exists a continuous differential form ∆ of degree h + 1 in
Rn such that
Z Z
∆∧ω = λ ∧ dω,
Rn Rn

for every Cc∞ differential form ω of degree n − h − 1 in Rn . Moreover let µ be a C 1

differential form of degree h + 1 in Rn and set E := {y ∈ Rn | ∆(y) = µ(y)}. Then
(dµ)|E (n+1) ≡ 0.

The second result is provided in Section 5. It originates from the following question: Is
it true that for every (measurable) function f : Rn → {0} ∪ [n, +∞] there exists E ⊂ Rn
such that dE = f almost everywhere? Some hasty considerations may mislead us into
thinking that the answer is yes, but on deeper reflection, it is not difficult to conclude
that the correct answer is no (cf. Example 5.1 below). However, somewhat surprisingly,
the following approximation property holds true.
Theorem. Let f : Rn → R be a measurable function such that f (x) ∈ {0} ∪ [n, +∞] for
a.e. x ∈ Rn . Then there exists a countable family {Fk }∞ n
k=1 of closed subsets of R such
n
that limk→∞ dFk (x) = f (x) for a.e. x ∈ R .

2. Preliminaries

2.1. General notation. The coordinates of Rn are denoted by (x1 , . . . , xn ) and we set
Di := ∂/∂xi . If k is any positive integer not exceeding n, then I(n, k) is the family of
integer multi-indices α = (α1 , . . . , αk ) such that 1 ≤ α1 < · · · < αk ≤ n. If α ∈ I(n, k),
then we denote by ᾱ the member of I(n, n−k) which complements α in {1, 2, . . . , n} in the
 SOME REMARKS ON POINTS OF LEBESGUE DENSITY AND DENSITY-DEGREE FUNCTIONS 3

natural increasing order (e.g., if α = (2, 3, 5) ∈ I(7, 3), then ᾱ = (1, 4, 6, 7) ∈ I(7, 4)). The
open ball of radius r centered at x ∈ Rn is denoted by B(x, r). Sometimes, for simplicity,
the sphere B(0, r) will be denoted by Br . Moreover χE is the characteristic function of
E. The equivalence relation of functions and the equivalence relation of subsets of Rn ,
with respect to the Lebesgue outer measure Ln , are both denoted by ∼. If E, F ⊂ Rn
and Ln (E \ F ) = 0, then we write E ⊂ e F (so that E ∼ F if and only if E ⊂ e F and F ⊂ e E).
c e c
Observe that E ⊂ F if and only if F ⊂ E .
e

2.2. Covectors and differential forms in Rn . If k is a positive integer not exceeding

n, then a k-covector (of Rn ) is a k-linear alternating map from (Rn )k to R. Let k (Rn )
V
V1 n
denote the set of all k-covectors. In particular (R ) is the dual space of Rn and we
Vk n
will denotethe
 standard dual basis by dx1 , . . . , dxn . The set (R ) is a vector space of
n
dimension k with the standard basis
(2.1) {dxα := dxα1 ∧ · · · ∧ dxαk | α ∈ I(n, k)}
where ∧ denotes the wedge product. Recall that k (Rn ) is equipped with the following
V

inner product naturally induced from Rn (making (2.1) an orthonormal basis):

X
hξ, ηi := ξα ηα
α∈I(n,k)

with X X
ξ= ξα dxα , η= ηα dxα .
α∈I(n,k) α∈I(n,k)

We observe that the following property holds (we set for simplicity dx := dx1 ∧ · · · ∧ dxn ):
^k ^n−k
(2.2) If ξ ∈ (Rn ) and hξ ∧ η, dxi = 0 for all η ∈ (Rn ), then ξ = 0.
Indeed, if ξ = α∈I(n,k) ξα dxα , then for all β ∈ I(n, k) we have 0 = hξ ∧ dxβ̄ , dxi =
P

ξβ hdxβ ∧ dxβ̄ , dxi, hence ξβ = 0.

A C H differential form of degree k (on Rn ) is a k-covector field

x ∈ Rn 7→
X
fα (x) dxα ,
α∈I(n,k)

with {fα | α ∈ I(n, k)} ⊂ C H (Rn ). A C H differential form of degree 0 is simply a function
in C H (Rn ). We recall that the addition and the exterior product of covectors naturally
induce the addition and the exterior product of differential forms. Moreover an exterior
derivative operator d is uniquely defined on C 1 differential forms and it holds that

• On C 1 differential forms of degree 0, the operator d agrees with the ordinary

differential on C 1 (R);
• d(λ + µ) = dλ + dµ, for all C 1 differential forms λ and µ of degree k;
• d(λ ∧ µ) = dλ ∧ µ + (−1)k λ ∧ dµ, for all C 1 differential forms λ and µ, if λ has
degree k;
• d(dλ) = 0, for all C 2 differential forms λ.
4 SILVANO DELLADIO

We also recall that, given a continuous differential form ω of degree n with compact
support and a measurable set E ⊂ Rn , the integral of ω on E is defined as follows:
Z Z
ω := hω(y), dx1 ∧ · · · ∧ dxn i dLn (y).
E E

For a comprehensive treatment of k-covectors and differential forms, we refer the reader
to the numerous books on differential geometry and geometric measure theory that deal
with this subject, e.g., [16] and [15].

2.3. Points of density. We recall the definition of m-density point (compare [2, 3, 4]).
Definition 2.1. Let m ∈ [n, +∞) and E ⊂ Rn . Then x ∈ Rn is said to be a “m-density
point of E” if
Ln (B(x, r)\E)
lim+ = 0.
r→0 rm
The set of m-density points of E is denoted by E (m) .
Remark 2.1. The following simple facts occur:

(1) Every interior point of E ⊂ Rn is an m-density point of E, for all m ∈ [n, +∞).
Thus, whenever E is open, one has E ⊂ E (m) for all m ∈ [n, +∞).
(2) If n ≤ m1 ≤ m2 < +∞ and E ⊂ Rn , then E (m2 ) ⊂ E (m1 ) . In particular, one has
E (m) ⊂ E (n) for all m ∈ [n, +∞).
(3) Let {Ej }j∈J be any family of subsets of Rn and m ∈ [n, +∞).
– One has
!(m) !(m)
\ \ (m) [ [ (m)
Ej ⊂ Ej , Ej ⊃ Ej
j∈J j∈J j∈J j∈J

– Let J be finite. Then

!(m)
\ \ (m)
(2.3) Ej = Ej ,
j∈J j∈J

while the identity

!(m)
[ [ (m)
Ej = Ej
j∈J j∈J

fails to be true in general, e.g., E1 = (−1, 0) and E2 = (0, 1) (with n = 1 and

J = {1, 2}).
– If J is countably infinite, then (2.3) can fail to be true, e.g., J = {j = 1, 2, . . . }
and Ej := B(0, 1/j).
(4) If E ⊂ Rn is measurable, then (E (n) )(m) = E (m) for all m ∈ [n, +∞). In particular
(E (n) )(n) = E (n) .
Proposition 2.1 ([6], Proposition 3.1). For all E ⊂ Rn , the set E (m) is measurable.
 SOME REMARKS ON POINTS OF LEBESGUE DENSITY AND DENSITY-DEGREE FUNCTIONS 5

Theorem 2.1 ([7], Corollary 4.1). If E is a measurable subset of Rn and m ∈ (n, +∞),
then
Ln (B(x, r) \ E)
( )
(m) n
E ∼ x ∈ R lim sup < +∞ .
r→0+ rm

The Lebesgue density theorem states that if E is a measurable subset of Rn , then almost
every x ∈ E is a n-density point of E. A remarkable family of sets that turn out to be
strictly more dense than generic measurable sets is that of finite perimeter sets. We recall
that the perimeter of a measurable set E ⊂ Rn , denoted by P (E), is the variation of χE ,
that is
(Z )
n
P (E) := sup div ϕ dL ϕ∈ Cc1 (Rn , Rn ), kϕk∞ ≤ 1 .
E

In the special case when ∂E is of class C 1 , the perimeter P (E) agrees with the natural
(n − 1)-dimensional hypersurface measure of ∂E. An excellent account of finite perimeter
sets can be found, e.g., in [14] and [1]. The following results show that the order of density
of every finite perimeter set is not less than the number
1
m0 := n + 1∗ = n + 1 +
n−1
and more precisely that m0 is the maximum order of density common to all sets of finite
perimeter.
Theorem 2.2 ([2], Lemma 4.1). Let E ⊂ Rn be measurable and such that P (E) < +∞.
Then E ∼ E (m0 ) .
Proposition 2.2 ([6], Proposition 4.1). For all m ∈ (m0 , +∞) there exists a closed set
Fm ⊂ Rn of positive measure and finite perimeter such that Fm(m) = ∅.

2.4. The density-degree function. Prior to providing the definition of the density-
degree function, observe that if E is a subset of Rn and x ∈ Rn , then the set {k ∈
[n, +∞) | x ∈ E (k) } is a (possibly empty) interval.
Definition 2.2. Let E be a subset of Rn . Then define the “density-degree function of E”
dE : Rn → [0, +∞] as follows
 n o
sup m ∈ [n, +∞) | x ∈ E (m) if x ∈ E (n)
dE (x) := 
0 6 E (n) .
if x ∈
When there exists k ∈ [n, +∞] such that
E ∼ d−1 n
E ({k}) = {x ∈ R | dE (x) = k}

we say that E is a “uniformly k-dense set”.

Example 2.1. If E is open, then E ⊂ d−1 E ({+∞}). Observe that strict inclusion can
occur, e.g. E := Br \ {0} (for which one has d−1
E ({+∞}) = Br ).
6 SILVANO DELLADIO

Example 2.2. Let m ∈ (2, +∞) and E be the set of points (x1 , x2 ) ∈ R2 satisfying
|x2 | > |x1 |m−1 . Since (as an elementary computation shows)
L2 (B(0, r) \ E)
lim+ ∈ (0, +∞),
r→0 rm
one has dE (0) = m and 0 6∈ E (m) .

This proposition collects a number of properties which have been proved in Proposition
5.1 and Proposition 5.2 of [7] (except (7) which is very easy to verify).
Proposition 2.3. Let E be a subset of Rn and m ∈ [n, +∞). The following properties
hold:

(1) The density-degree function dE is measurable.

(2) d−1 −1
E ({k}) ∩ dE ({m}) = ∅, if k 6= m (k ≥ n).

(3) If E is measurable, then the set

n o
m ∈ (n, +∞) | Ln (d−1
E ({m})) > 0

is at most countable.
(4) d−1 −1
E ((m, +∞]) = dE ((m, kdE k∞ ]) = E (k) .
S
k>m

(5) If m > n then

d−1 −1
E (l) ,
\
E ([m, +∞]) = dE ([m, kdE k∞ ]) =
l∈[n,m)

while
d−1 −1
E ([n, +∞]) = dE ([n, kdE k∞ ]) = E
(n)
.
(6) d−1
E ((m, +∞]) ⊂ E
(m)
⊂ d−1
E ([m, +∞]).

(7) Let E (m) 6= ∅. Then dE |E (m) ≡ m if and only if kdE k∞ = m.

Remark 2.2. Both the inclusions in statement (4) of Proposition 2.3 may be strict (cf.
Proposition 2.4 below and Example 2.2, respectively).
Remark 2.3. Let Ω and E be, respectively, an open subset of Rn and a measurable subset
of Ω. We observe that the existence of even a single point x ∈ Ω such that dE (x) < +∞
yields Ln (Ω \ E) > 0. This simple observation might lead us to believe that Ln (E) must
be small if the set of such points x has a large measure. Proposition 2.4 and Theorem 2.3
below show, in particular, that this is not true.

The following result establishes that a bounded open set in Rn can be arbitrarily appro-
ximated from inside by closed uniformly n-dense sets.
Proposition 2.4 ([7], Proposition 5.4). Let Ω be a bounded open subset of Rn . Then
for all C < Ln (Ω) there exists an uniformly n-dense closed subset F of Ω such that
Ln (F ) > C.
 SOME REMARKS ON POINTS OF LEBESGUE DENSITY AND DENSITY-DEGREE FUNCTIONS 7

We expect that Proposition 2.4 can be extended to a result of approximation from inside
by closed uniformly k-dense sets, for all k ≥ n. We are not yet able to resolve this
conjecture, but we have the following result.
Theorem 2.3 ([7], Theorem 5.1). Let Ω be a bounded open subset of Rn and let m ∈
(n, +∞). Then for all C < Ln (Ω) and for all t ∈ (n, m) there exist a closed subset F of
Ω and an open subset U of Ω such that:

(1) d−1 n n n
F ([t, +∞]) ⊃ Ω \ U and L (U ) < L (Ω) − C (hence F ⊃ Ω \ U and L (F ) > C);

(2) One has F (m) = ∅ (hence kdF k∞ ≤ m).

In particular, one has t ≤ dF (x) ≤ m for all x ∈ Ω \ U .

Remark 2.4. When ∂Ω is Lipschitz, it is obvious that the closed set F in Proposition
2.4 and in Theorem 2.3 can be chosen so that F ⊂ Ω.

Finally, observe that Theorem 2.2 can be restated as follows:

Proposition 2.5 ([7], Proposition 5.3). Let E ⊂ Rn be measurable and such that P (E) <
+∞. Then one has d−1E ([m0 , +∞]) ∼ E.

3. A simple characterization of superdensity points

Proposition 3.1. Let E ⊂ Rn be measurable, x ∈ Rn and m ∈ [n, +∞).

(1) If x ∈ E (m) , then

y−x
Z  
(3.1) g(y)ϕ dLn (y) = o(rm ) (as r → 0+)
Ec r
for all ϕ ∈ Cc (Rn ) and for every measurable function g : Rn → R which is bounded
in a neighborhood of x.
(2) Let g ∈ C(Rn ) be such that g(x) 6= 0 and (3.1) holds for all ϕ ∈ Cc (Rn ). Then
x ∈ E (m) .

Proof. (1) Let us consider x ∈ E (m) , ϕ ∈ Cc (Rn ) and an arbitrary measurable function
g : Rn → R which is bounded in a neighborhood of x. If R is a positive number such that
supp(ϕ) ⊂ B(0, R), then
y−x y−x
Z   Z  
n
g(y)ϕ dL (y) = g(y)ϕ dLn (y)
E c r B(x,rR)∩E c r
!
≤ sup |g| kϕk∞ Ln (B(x, rR) ∩ E c ).
B(x,rR)
8 SILVANO DELLADIO

Hence (3.1) follows at once.

(2) Let us consider ϕ ∈ Cc (Rn ) such that

ϕ(Rn ) = [0, 1], supp(ϕ) ⊂ B(0, 2), ϕ|B(0,1) = 1.
Moreover, without loss of generality, we can assume that g is positive in a neighborhood
of x. Hence two positive constants p and r0 have to exist such that inf B(x,2r0 ) g ≥ p. For
all r ∈ (0, r0 ] we have
Z
p Ln (B(x, r) ∩ E c ) ≤ g(y) dLn (y)
B(x,r)∩E c
!
Z
y−x
≤ g(y) ϕ dLn (y),
E c r
hence the conclusion follows from (3.1). 

This corollary is a trivial consequence of Proposition 3.1.

Corollary 3.1. Let E ⊂ Rn be measurable, x ∈ Rn and m ∈ [n, +∞). Then each of the
following properties is equivalent to x ∈ E (m) :

(1) The equation

Z Z
n
E−x
ϕ dL = ϕ dLn + o(rm−n ) (as r → 0+)
r
Rn

holds for all ϕ ∈ Cc (Rn ).

(2) There exists g ∈ C(Rn ) such that g(x) 6= 0 and (3.1) holds for all ϕ ∈ Cc (Rn ).
(3) The identity (3.1) holds for all ϕ ∈ Cc (Rn ) and for every measurable function
g : Rn → R which is bounded in a neighborhood of x.

The next result is also a very easy consequence of Proposition 3.1.

Corollary 3.2. Let us consider a measurable set E ⊂ Rn , x ∈ E (m) with m ∈ [n, +∞)
and a measurable function Γ : Rn → R which is continuous at x. Assume that there exists
ψ ∈ Cc (Rn ) such that Rn ψ dLn 6= 0 and (for r small enough)
R

k Z
y−x y−x
Z    
n n−m
dLn (y)
X
Γ(y)ψ dL (y) ≤ r gi (y)ϕi
R n r i=1 E
c r
where g1 , . . . , gk : Rn → R is a family of measurable functions which are bounded in a
neighborhood of x and ϕ1 , . . . , ϕk ∈ Cc (Rn ). Then Γ(x) = 0.

Proof. From (3.1) it follows that

y−x
Z  
Γ(y)ψ dLn (y) = rn−m o(rm ) = o(rn ) (as r → 0+).
Rn r
 SOME REMARKS ON POINTS OF LEBESGUE DENSITY AND DENSITY-DEGREE FUNCTIONS 9

On the other hand, we have also

y−x
Z   Z
n n
Γ(y)ψ dL (y) = r Γ(x + rz)ψ (z) dLn (z),
Rn r Rn

so that
Z
Γ(x + rz)ψ (z) dLn (z) = r−n o(rn ) (as r → 0+).
Rn

Hence (letting r → 0+)

Z
Γ(x) ψ dLn = 0
Rn

ψ dLn 6= 0.
R
and the conclusion follows recalling that Rn 

We are interested in Corollary 3.2 as it provides a common argument for the proofs of
several theorems we have obtained in our work on superdensity, e.g., [2, Theorem 2.1]
(the oldest one) and [10, Theorem 3.6] (the most recent one). Having it in mind can be
useful for yielding new applications. As an example, let us prove the following result.

Theorem 3.1. Let λ be a continuous differential form of degree h on Rn (with h ≥ 0)

having the following property: There exists a continuous differential form ∆ of degree h+1
in Rn such that

Z Z
(3.2) ∆∧ω = λ ∧ dω,
Rn Rn

for every Cc∞ differential form ω of degree n − h − 1 in Rn . Moreover let µ be a C 1

differential form of degree h + 1 in Rn and set E := {y ∈ Rn | ∆(y) = µ(y)}. Then
(dµ)|E (n+1) ≡ 0.

Proof. Let x ∈ E (n+1) . Consider an arbitrary C ∞ differential form θ of degree n − h − 2

in Rn and define Γ : Rn → R as

Γ(y) := hdµ(y) ∧ θ(y), dx1 ∧ · · · ∧ dxn i (y ∈ Rn ).

10 SILVANO DELLADIO

Moreover let ψ ∈ Cc∞ (Rn ) be such that ψdLn 6= 0. Then, from the Stokes theorem,
R
Rn
we obtain (for r small enough)
y−x y−x
Z   Z  
Γ(y)ψ dLn (y) = ψ dµ(y) ∧ θ(y)
Rn r Rn r
y−x
Z   
=− d ψ ∧ µ(y) ∧ θ(y)
Rn r
y−x
Z  
h
+ (−1) ψ µ(y) ∧ dθ(y)
Rn r 
y−x
Z  
=− d ψ ∧ ∆(y) ∧ θ(y)
E r 
y−x
Z  
− d ψ ∧ µ(y) ∧ θ(y)
E c r
y−x
Z  
+ (−1)h ψ ∆(y) ∧ dθ(y)
E r 
y−x
Z 
+ (−1)h ψ µ(y) ∧ dθ(y)
Ec r
y−x
Z   
=− d ψ ∧ ∆(y) ∧ θ(y)
Rn r
y−x
Z   
+ d ψ ∧ [∆(y) − µ(y)] ∧ θ(y)
Ec r
y−x
Z  
+ (−1)h ψ ∆(y) ∧ dθ(y)
Rn r
y−x
Z  
h
+ (−1) ψ [µ(y) − ∆(y)] ∧ dθ(y).
Ec r
Hence
y−x
Z  
(3.3) Γ(y)ψ dLn (y) = I(r) + J(r),
Rn r
where
y−x y−x
Z     
I(r) := −d ψ ∧ ∆(y) ∧ θ(y) + (−1)h ψ ∆(y) ∧ dθ(y)
Rn r r
and
y−x y−x
Z     
h
J(r) := d ψ ∧ [∆(y) − µ(y)] ∧ θ(y) + (−1) ψ [µ(y) − ∆(y)] ∧ dθ(y).
Ec r r
Now observe that
y−x y−x
Z       
h
(−1) I(r) = ∆(y) ∧ d ψ ∧ θ(y) + ψ dθ(y)
Rn r  r
y−x
Z   
= ∆(y) ∧ d ψ θ(y) ,
Rn r
hence
(3.4) I(r) = 0,
 SOME REMARKS ON POINTS OF LEBESGUE DENSITY AND DENSITY-DEGREE FUNCTIONS 11

by assumption (3.2). Moreover, for all r ∈ (0, 1], one has

n Z
1X y−x
 
|J(r)| = (Di ψ) dxi ∧ [∆(y) − µ(y)] ∧ θ(y)
r i=1 E c r
y−x
Z  
(3.5) + (−1)h ψ [µ(y) − ∆(y)] ∧ dθ(y)
Ec r
1 n+1 y−x
X Z  
≤ gi (y)ϕi dLn (y)
r i=1 E c r
with
ϕi := Di ψ, gi := hdxi ∧ (∆ − µ) ∧ θ, dx1 ∧ · · · ∧ dxn i (i = 1, . . . , n)
and
ϕn+1 := ψ, gn+1 := h(µ − ∆) ∧ dθ, dx1 ∧ · · · ∧ dxn i.
From (3.3), (3.4), (3.5) and Corollary 3.2 (with m = k = n + 1) we obtain Γ(x) = 0. The
conclusion follows from the arbitrariness of θ, by recalling (2.2). 
Remark 3.1. Let λ and ω be, respectively, a C 1 differential form of degree h and a Cc1
differential form of degree n − h − 1 on Rn . Since d(λ ∧ ω) = dλ ∧ ω + (−1)h λ ∧ dω, we
have Z Z Z Z
λ ∧ dω = (−1)h d(λ ∧ ω) + (−1)h+1 dλ ∧ ω = (−1)h+1 dλ ∧ ω.
Rn Rn Rn Rn
So, when λ is of class C 1 , the condition (3.2) is verified with ∆ = (−1)h+1 dλ. Hence
Theorem 3.1 yields the following result (provided in [2, Theorem 2.1]): Let λ and µ be
C 1 differential forms on Rn of degree h and h + 1, respectively (with h ≥ 0). If define
E := {y ∈ Rn | dλ(y) = µ(y)}, then (dµ)|E (n+1) ≡ 0.

4. Some further properties of the density-degree functions

Proposition 4.1. Given two subsets E, F of Rn , the following properties hold:

(1) If E ∼ F , then E (m) = F (m) for all m ∈ [n, +∞), hence dE = dF .

(2) If E, F are measurable and E 6∼ F , then {dE 6= dF } has positive measure.

Proof. (1) Since Ln (E ∩F c ) = Ln (F ∩E c ) = 0, the sets E ∩F c and F ∩E c are measurable.

It follows that (for all x ∈ Rn and r > 0)
Ln (B(x, r) \ E) = Ln (B(x, r) ∩ [(E ∩ F ) ∪ (E ∩ F c )]c )
= Ln (B(x, r) ∩ (E ∩ F )c ∩ (E ∩ F c )c )
= Ln (B(x, r) ∩ (E ∩ F )c ) − Ln (B(x, r) ∩ (E ∩ F )c ∩ (E ∩ F c ))
where the second term on the right above is 0 by monotonicity. Since the last term is
symmetric in E, F , one has
Ln (B(x, r) \ F ) = Ln (B(x, r) \ E) = Ln (B(x, r) ∩ (E ∩ F )c ),
12 SILVANO DELLADIO

hence E (m) = F (m) for all m ≥ n.

(2) First of all, recall that (by a well-known result on the Lebesgue set, e.g., see Corollary
1.5 in [18, Chapter 3]) a measure zero set N ⊂ Rn has to exist such that

(4.1) E \ N ⊂ E (n) , (F c \ N ) ∩ F (n) = ∅

and

(4.2) F \ N ⊂ F (n) , (E c \ N ) ∩ E (n) = ∅.

Moreover, by hypothesis, at least one of the following inequalities must hold:

Ln (E ∩ F c ) > 0, Ln (F ∩ E c ) > 0.

• If the first inequality holds, then, by (4.1), for all x in the positive measure set
E ∩ F c ∩ N c one has dE (x) ≥ n (in that x ∈ E ∩ N c ⊂ E (n) ) and dF (x) = 0 (in
that x ∈ F c ∩ N c , hence x 6∈ F (n) ).
• If instead the second inequality holds, then, by (4.2), for all x in the positive
measure set F ∩ E c ∩ N c one has dF (x) ≥ n (in that x ∈ F ∩ N c ⊂ F (n) ) and
dE (x) = 0 (in that x ∈ E c ∩ N c , hence x 6∈ E (n) ).

The following property follows immediately from (1) of Proposition 4.1, by also recalling
that Lebesgue outer measure is Borel regular.

Corollary 4.1. For every measurable subset E of Rn there exists a Borel set B such that
E ⊂ B, Ln (E) = Ln (B) and dE = dB .

Proposition 4.2. Let {Ej }j∈J be any family of subsets of Rn . Then the following ine-
qualities hold:

(4.3) d∩j∈J Ej ≤ inf dEj , d∪j∈J Ej ≥ sup dEj .

j∈J j∈J

If J is finite, then the first one turns into the equality d∩j∈J Ej = minj∈J dEj , while the
identity d∪j∈J Ej = maxj∈J dEj fails to be true in general.

Proof. For all x 6∈ (∩j∈J Ej )(n) we obviously have

d∩j∈J Ej (x) = 0 ≤ inf dEj (x).

j∈J
 SOME REMARKS ON POINTS OF LEBESGUE DENSITY AND DENSITY-DEGREE FUNCTIONS 13

So let us suppose x ∈ (∩j∈J Ej )(n) . Then the set {k ≥ n | x ∈ (∩j∈J Ej )(k) } is non-empty
and
d∩j∈J Ej (x) = sup{k ≥ n | x ∈ (∩j∈J Ej )(k) }
(k)
≤ sup{k ≥ n | x ∈ ∩j∈J Ej }
(k)
= inf sup{k ≥ n | x ∈ Ej }
j∈J

= inf dEj (x),

j∈J

where equality holds in the second line whenever J is finite (cf. Remark 2.1). This
concludes the proof of the part concerning d∩j∈J Ej .

(n)
For all x 6∈ (∪j∈J Ej )(n) , one also has x 6∈ Ej for all j ∈ J (cf. Remark 2.1) and thus
d∪j∈J Ej (x) = sup dEj (x) = 0.
j∈J

On the other hand, if x ∈ (∪j∈J Ej )(n) , then the set {k ≥ n | x ∈ (∪j∈J Ej )(k) } is non-empty
and (cf. Remark 2.1)
d∪j∈J Ej (x) = sup{k ≥ n | x ∈ (∪j∈J Ej )(k) }
(k)
≥ sup{k ≥ n | x ∈ ∪j∈J Ej }
(k)
= sup sup{k ≥ n | x ∈ Ej }
j∈J

= sup dEj (x).

j∈J

To verify the last assertion, we can consider the example provided in Remark 2.1 (n = 1,
J = {1, 2}, E1 = (−1, 0), E2 = (0, 1)) which yields
d∪j∈J Ej = +∞χ(−1,1) , max dEj = +∞χ(−1,1)\{0} .
j∈J

5. Does the set having an arbitrarily prescribed

density-degree function always exist?

As we have stated in Proposition 2.3, every density-degree function (relative to a subset

of Rn ) is measurable and takes its values in {0} ∪ [n, +∞]. Rough considerations on this
subject could lead us to believe that for every measurable function f : Rn → R such that
f (x) ∈ {0} ∪ [n, +∞] at a.e. x ∈ Rn there exists E ⊂ Rn for which dE ∼ f . The following
example disproves this.
Example 5.1. Let F be a measurable subset of Rn such that 0 < kdF k∞ < +∞ (observe
that F can be provided by Proposition 2.4 or Theorem 2.3) and let m ∈ (kdF k∞ , +∞).
Then consider the measurable function f := mχF and prove (by contradiction) that there
is no measurable set E ⊂ Rn for which dE ∼ f . In fact, if such a set E exists, then
14 SILVANO DELLADIO

E ∼ E (n) ∼ F (cf. Definition 2.2), hence dF = dE ∼ mχF (cf. Proposition 4.1) and this
contradicts the hypothesis m > kdF k∞ .

These limitations, however, are not sufficient to prevent good approximation properties.
In fact, the following result holds.
Theorem 5.1. Let f : Rn → R be a measurable function such that f (x) ∈ {0} ∪ [n, +∞]
for a.e. x ∈ Rn . Then there exists a countable family {Fk }∞
k=1 of closed subsets of R
n
n
such that limk→∞ dFk (x) = f (x) for a.e. x ∈ R .

Proof. Let N := f −1 ([−∞, 0) ∪ (0, n)) and define


f (x) if x ∈ Rn \ N
fe(x) :=
0 if x ∈ N .

Note that fe is equivalent to f (since L(N ) = 0, by assumption) and takes all of its values
in {0} ∪ [n, +∞]. By [17, Theorem 1.17] there exists a nondecreasing sequence of simple
measurable functions on Rn
Mk
X
sk = ak,j χEk,j (k = 1, 2, . . . )
j=1

with the following properties:

(i) {Ek,j }M e−1 ([n, +∞]) and {a }Mk ⊂ [n, +∞);

j=1 is a measurable partition of E := f
k
k,j j=1
n
(ii) limk→∞ sk (x) = f (x), for all x ∈ R .
e

Now arguing as in the proof of [18, Theorem 4.3], with

Mk
X
ϕk := sk χBk = ak,j χEk,j ∩Bk (k = 1, 2, . . . ),
j=1

we can find a sequence of simple functions

Nk
X
ψk = bk,j χRk,j (k = 1, 2, . . . ),
j=1

where {Rk,j }N Nk
j=1 is a collection of disjoint closed rectangles and {bk,j }j=1 ⊂ [n, +∞), such
k

that
(5.1) lim ψk (x) = fe(x) = f (x), for a.e. x ∈ Rn .
k→∞

Without loss of generality we can also assume bk,j ∈ (n, +∞) for all k, j (it is enough
to replace bk,j with bk,j + 1/k), hence a sequence of positive real numbers {εk }∞
k=1 has to
exist such that limk→∞ εk = 0 and
bk,j − εk > n, for all j = 1, . . . , Nk .
 SOME REMARKS ON POINTS OF LEBESGUE DENSITY AND DENSITY-DEGREE FUNCTIONS 15

Now let Ωk,j denote the interior of Rk,j . By applying Theorem 2.3 and Remark 2.4, we
find a closed set Fk,j ⊂ Ωk,j and an open set Uk,j ⊂ Ωk,j satisfying
2−k 2−k
(5.2) Ln (Uk,j ) < , Ln (Fk,j ) > Ln (Ωk,j ) −
Nk Nk
and
|dFk,j (x) − bk,j | < εk , for all x ∈ Ωk,j \ Uk,j .
In particular, if define
N
[k N
[k N
[k
(5.3) Ωk := Ωk,j , Fk := Fk,j , Uk := Uk,j ,
j=1 j=1 j=1

then we get
(5.4) |dFk − ψk | < εk in Ωk \ Uk .
Also one obviously has
N
[k
(5.5) dFk = ψk = 0 in Rkc , where Rk := Rk,j .
j=1

Observe that
∞ [ ∞
! !
\ [
W := Uk ∪ ∂Rh
l=1 k>l h=1

is a set of measure zero, by (5.2) and (5.3). Moreover, if

∞ \
! ∞
!c ∞
" ! ∞
!c #
x ∈ Wc = Ukc ∩ Ukc ∩
[ [ [ \ [
∂Rh = ∂Rh ,
l=1 k>l h=1 l=1 k>l h=1

then there exists lx ≥ 1 such that

! ∞
!c
Ukc
\ [
x∈ ∩ ∂Rh ,
k>lx h=1

namely
∞
!c
Ukc ⊂ (Ωk \ Uk ) ∪ Rkc , for all k > lx .
[
x∈ ∩ ∂Rh
h=1

Hence, by (5.4) and (5.5), we get

|dFk (x) − ψk (x)| < εk for all k > lx .
The conclusion follows from (5.1). 
Corollary 5.1. Let E be a measurable subset of Rn . Then there exists an increasing
sequence {Fl }∞ n
l=1 of uniformly n-dense closed subsets of R such that ∪l Fl ∼ E.
16 SILVANO DELLADIO

Proof. By applying Theorem 5.1 to the measurable function f := nχE , we find a countable
family {Fek }∞ n n
k=1 of closed subsets of R and a measure zero set N ⊂ R such that

(5.6) lim dFek (x) = n, for all x ∈ E \ N

k→∞

and
(5.7) lim dFek (x) = 0, for all x ∈ E c \ N .
k→∞

From (5.6) and (5.7) we get, respectively (cf. Definition 2.2),

(n) (n)
Ec \ N ⊂ (Fek )c .
[\ [\
(5.8) E\N ⊂ Fek ,
l k≥l l k≥l

From the second inclusion in (5.8) it follows that

[\ (n) \[ (n)
Fek ⊂ Fek ⊂
e E.
l k≥l l k≥l

Hence and by the first inclusion in (5.8) we obtain

[\ (n) [ \
E∼ Fek ∼ Fl , with Fl := Fek .
l k≥l l k≥l

Observe that {Fl }∞

is an increasing sequence of closed sets. The only thing left to
l=1
prove is that each Fl is uniformly n-dense. For this purpose, observe first that for all
(n)
x ∈ Fl ∩ E \ N we have
n ≤ dFl (x) ≤ inf dFek (x) = n,
k≥l

by Definition 2.2, Proposition 4.2 and (5.6). Now the conclusion follows from the equiva-
(n)
lence Fl ∩ E \ N ∼ Fl . 

References

[1] L. Ambrosio, N. Fusco, D. Pallara: Functions of bounded variation and free discontinuity problems.
Oxford Mathematical Monographs, Oxford University Press Inc. 2000.
[2] S. Delladio: Functions of class C 1 subject to a Legendre condition in an enhanced density set. Rev.
Mat. Iberoam. 28 (2012), no. 1, 127-140.
[3] S. Delladio: A short note on enhanced density sets. Glasg. Math. J. 53 (2011), no. 3, 631-635.
[4] S. Delladio: A Whitney-type result about rectifiability of graphs. Riv. Math. Univ. Parma (N.S.) 5
(2014), no. 2, 387-397.
[5] S. Delladio: Some new results about the geometry of sets of finite perimeter. Proc. Royal Soc. Edin-
burgh 146A (2016), 79-105.
[6] S. Delladio: A note on some topological properties of sets with finite perimeter. Glasg. Math. J., 58
(2016), no. 3, 637-647.
[7] S. Delladio: Density-degree function for subsets of Rn . Houston J. Math. 45 (2019), n. 3, 743-762.
[8] S. Delladio: The tangency of a C 1 smooth submanifold with respect to a non-involutive C 1 distribution
has no superdensity points. Indiana Univ. Math. J. 68 (2019), n. 2, 393-412.
[9] S. Delladio: Approximate continuity and differentiability with respect to density degree. An applica-
tion to BV and Sobolev functions. Accepted by Anal. Math.
 SOME REMARKS ON POINTS OF LEBESGUE DENSITY AND DENSITY-DEGREE FUNCTIONS 17

[10] S. Delladio: The identity G(D)f = F for a linear partial differential operator G(D). Lusin type and
structure results in the non-integrable case. (available on ResearchGate) Proc. R. Soc. Edinburgh,
151(2021), issue 6, 1893-1919.
[11] S. Delladio: Some results about the structure of primitivity domains for linear partial differential
operators with constant coefficients. Accepted by Mediterr. J. Math. (available on ResearchGate).
[12] L.C. Evans, R.F. Gariepy: Measure Theory and Fine Properties of Functions. (Studies in Advanced
Math.) CRC Press 1992.
[13] H. Federer: Geometric Measure Theory. Springer-Verlag 1969.
[14] E. Giusti: Minimal surfaces and functions of bounded variation. Monographs in Mathematics 80,
Birkhäuser 1984.
[15] S.G. Krantz, H.R. Parks: Geometric integration theory. Cornerstones, Birkhäuser 2008.
[16] J.M. Lee: Introduction to smooth manifolds (second edition). Graduate Texts in Mathematics 218,
Springer Verlag 2013.
[17] W. Rudin: Real and Complex Analysis. McGraw-Hill 1970.
[18] R. Shakarchi and E.M. Stein: Real analysis. Measure theory, integration and Hilbert spaces. Prince-
ton Lectures in Analysis III. Princeton University Press, Princeton and Oxford 2005.