CARDINAL FUNCTIONS OF THE HYPERSPACE OF

CONVERGENT SEQUENCES

DAVID MAYA*, PATRICIA PELLICER-COVARRUBIAS**,

AND ROBERTO PICHARDO-MENDOZA***

Abstract. The symbol Sc (X) denotes the hyperspace of all nontrivial

convergent sequences in a Hausdorff space X. This hyperspace is endowed
with the Vietoris topology. In the current paper, we compare the cellularity,
the tightness, the extent, the dispersion character, the net weight, the i-
weight, the π-weight, the π-character, the pseudocharacter and the Lindelöf
number of Sc (X) with the corresponding cardinal function of X. We also
answer a question posed by the authors in a previous paper.

1. Introduction
Convergence of sequences is an important tool to determine topological prop-
erties in Hausdorff spaces. On the other hand, the study of hyperspaces can
provide information about the topological behavior of the original space and
vice versa. In connection with both concepts, the hyperspace consisting of all
nontrivial convergent sequences Sc (X), of a metric space X without isolated
points, was introduced in [3]. Interesting properties of this hyperspace are pre-
sented in [6] where the study was extended to Hausdorff spaces; among other
results, in [6, Section 6], three main cardinal functions (weight, character and
density) of a space X are compared with those of Sc (X). Now, we will do a
similar analysis for several other cardinal functions. The purpose of the current
paper is to present interrelations between several cardinal functions of Sc (X)
and the corresponding cardinal functions of X; namely, we will consider cel-
lularity, tightness, extent, dispersion character, net weight, i-weight, π-weight,
π-character, pseudocharacter and Lindelöf number. We also answer a question
posed in [6, Question 6.9].

2010 Mathematics Subject Classification. 54A20, 54A25, 54B20.

Key words and phrases. Hyperspace of nontrivial convergent sequences, cellularity, tight-
ness, extent, dispersion character, net weight, i-weight, π-weight, π-character, pseudocharacter,
Lindelöf number.
The research of the first author was supported by Programa de Becas Posdoctorales en la
UNAM, 2015-2016.
1
2 D. MAYA, P. PELLICER-COVARRUBIAS, AND R. PICHARDO-MENDOZA

2. Preliminaries
All topological notions and all set-theoretic notions whose definition is not
included here should be understood as in [1] and [5], respectively.
The symbol ω denotes both, the first infinite ordinal and the first infinite
cardinal. In particular, we consider all nonnegative integers as ordinals too;
thus, n ∈ ω implies that n = {0, . . . , n − 1} and ω \ n = {k ∈ ω : k ≥ n}. The
set ω \ {0} is denoted by N. The successor of an infinite cardinal κ is the ordinal
κ + 1 = κ ∪ {κ} and so the symbols i ∈ κ + 1, i < κ + 1 and i ≤ κ all represent
the same. As usual, c will be used to represent the cardinality of the real line,
R.
If X is a set and κ is a cardinal, |X| will represent the cardinality of X and
the symbols [X]κ , [X]≤κ and [X]<κ denote the families of all subsets of X whose
cardinality is κ, ≤ κ and < κ, respectively. In particular, [X]<ω is the collection
of all finite subsets of X, [X]ω is the collection of all infinite countable subsets of
X and [X]<n+1 is the collection of all subsets of X having at most n elements,
whenever n ∈ ω.
For a function f , ran(f ) will denote its range and dom(f ) will represent its
domain. Given a subset A of dom(f ), the set {f (x) : x ∈ A} will be designated
by the symbol f [A].
The cartesian product S of a family {Xα : α ∈ I} of sets, i. e., the set of all
functions
Q f from I into α∈I Xα such that f (α) ∈ Xα for every α ∈ I, is denoted
by α∈I Xα .
In this paper, space means Hausdorff space. A completely regular space is
called a Tychonoff space. For a space X, the symbol τX will denote the collection
of all open subsets of X. Also, for a set A ⊆ X, we will use intX A and clX A (or,
if there is no risk of confusion, A) to represent its interior in X and its closure
in X, respectively.
The topological product of a family of topological spaces {Xα : αQ∈ I} is
the topological space which results of endowing the cartesian product α∈I Xα
with the product topology.
A convergent sequence in a topological space X is a function f from ω into X
for which there is x ∈ X in such a way that: for each U ∈ τX with x ∈ U there
exists n ∈ ω with f [ω \ n] ⊆ U . In this case, we will say either that f converges
to x or x is the limit of f , and this fact will be denoted by either lim f (n) = x
n→∞
or f (n) → x. We shall write (f (n))n∈ω to refer to f . If | ran(f )| = ω, we say
that f is nontrivial. In connection with this concept, in this paper, a subset S of
a space X will be called a nontrivial convergent sequence in X if: S ∈ [X]ω and
there is x ∈ S in such a way that S \ U ∈ [X]<ω for each U ∈ τX with x ∈ U (see
[6]). When this happens, the point x is called the limit point of S and we will
say that S converges to x and write either S → x or lim S = x. Throughout this
 CARDINAL FUNCTIONS OF Sc (X) 3

paper, the reader will be able to identify from the context what is the intended
meaning of nontrivial convergent sequence in the discussion.
For a space X, let

C L (X) = {A ⊆ X : A is closed in X and A 6= ∅},

K (X) = {A ∈ C L (X) : A is compact} and
Sc (X) = {S ∈ K (X) : S is a nontrivial convergent sequence in X}.

Given a family U of subsets of X, we define

n [ o
hU i = A ∈ C L (X) : A ⊆ U ∧ ∀ U ∈ U (A ∩ U 6= ∅) .

The Vietoris topology is the topology on C L (X) generated by the base con-
sisting of all sets of the form hU i, where U ∈ [τX ]<ω (see [7, Proposition 2.1,
p. 155]). The hyperspaces K (X) and Sc (X) will be considered as subspaces of
C L (X). In particular, a base for the topology of Sc (X) consists of all sets of
the form hU ic = hU i ∩ Sc (X), where U ∈ [τX ]<ω . For each n ∈ N, we will
denote by Fn (X) the subspace [X]<n+1 \ {∅} of C L (X).
For a subset U of a space X, let U + = h{U }i, Uc+ = h{U }ic , U − = h{X, U }i
and Uc− = h{X, U }ic . Thus, when V is open (closed, resp.) in X, then V +
and V − are open (closed, resp.) in C L (X), and hence, Vc+ and Vc− are open
(closed, resp.) in Sc (X).
For the sake of simplicity, we will adopt the following convention: if S is a
nontrivial convergent sequence in a space X, we will say that {xn : n ≤ ω} is
an adequate enumeration of S provided that S = {xn : n ≤ ω}, lim S = xω and
xi 6= xj whenever i < j ≤ ω.
A cellular family in a topological space X is a pairwise disjoint family of
nonempty open subsets of X. The collection of all finite cellular families of X
is denoted by C(X).
A topological space having no isolated points will be called crowded. For a
space X and x ∈ X, we will denote by Sc (X, x) the set {S ∈ Sc (X) : lim S = x}
and define LX = {y ∈ X : Sc (X, y) 6= ∅}.
In [6], the following concept was introduced. We will say that a topological
space X is abundant in sequences if LX is dense in X. Equivalently, a space X
is abundant in sequences if and only if Sc (X) is dense in C L (X). Observe that
each space which is abundant in sequences is crowded.
Finally, given an infinite cardinal κ, endow κ with the discrete topology and
set J(κ) = ([0, 1]×κ)/({0}×κ), i. e., J(κ) is the quotient space of the topological
product [0, 1] × κ which results of collapsing the set {0} × κ to a single point.
Similarly, define S(κ) = ((ω + 1) × κ)/({ω} × κ), where the ordinal ω + 1 is
considered as a linearly ordered topological space. Usually, J(κ) and S(κ) are
called the hedgehog of κ spines and the sequential fan of κ spines, respectively.
4 D. MAYA, P. PELLICER-COVARRUBIAS, AND R. PICHARDO-MENDOZA

We will use the acronym LOTS instead of the phrase “linearly ordered topo-
logical space”.

3. Auxiliary results
Lemma 3.1 ([7, Theorem 4.9, p. 163]). If X is a Ti -space for i ∈ {2, 3, 3 21 },
then K (X) is Ti and therefore so is Sc (X).
Proposition 3.2 ([6, Proposition 3.2]). For an arbitrary space X, {hU ic : U ∈
C(X)} is a base for Sc (X).
The weight of a space X, w(X), is the least cardinality of a base for X.
Lemma 3.3 ([6, Theorem 6.5]). Let X be a space with more than one point.
Then, w(X) = w(Sc (X)) if and only if Sc (X) 6= ∅.
Lemma 3.4. Let X be a space, let x ∈ X and let U ∈ τX be such that x ∈ U .
If x ∈ LX , then Uc+ ∩ Sc (X, x) 6= ∅.
Proof. Since x ∈ LX , there exists Q ∈ Sc (X, x). Then, Q\U is finite. Therefore,
S = Q ∩ U ∈ Uc+ and lim S = x. 
Proposition 3.5. For an arbitrary space X, LSc (X) = Sc (X). In particular,
Sc (X) is crowded and, when Sc (X) 6= ∅, Sc (X) is infinite.
Proof. Let S ∈ Sc (X) and let {xn : n ≤ ω} be an adequate enumeration of
S. For each n < ω, let Sn = S \ {xn }. Observe that {Sn : n < ω} ∪ {S} ∈
Sc (Sc (X), S). Thus, S ∈ LSc (X) . 
Recall that a subset of a topological space is nowhere dense if the interior of
its closure is empty.
Lemma 3.6 ([6, Lemma 3.4]). Let X be a space, let A ∈ [C L (X)] <ω
be a
pairwise disjoint family and let n ∈ N. If S ∈ Sc (X) satisfies that S ∩ A = ∅,
S
then ( )
Y
En (A , S) = S ∪ ran(t) : t ∈ Fn (A)
A∈A
is a closed nowhere dense subset of Sc (X) which is homeomorphic to Fn (A).
Q
A∈A

4. Cellularity
The least cardinal κ such that every cellular family of a topological space X
has cardinality ≤ κ is called the cellularity, or the Suslin number, of X and is
denoted by c(X).
Lemma 4.1. Let X be a space. If |X| ≥ ω, then c(X) ≥ ω. In particular, if
Sc (X) 6= ∅, then c(Sc (X)) ≥ ω.
 CARDINAL FUNCTIONS OF Sc (X) 5

Proof. Fix n < ω and let F ∈ [X]n+1 . Then, there exists U ∈ C(X) satisfying
that |U | = n + 1 and that each element of U intersects F . This shows that
c(X) > n for every n < ω. Therefore, c(X) ≥ ω.
Finally, assume that Sc (X) 6= ∅. Proposition 3.5 implies that |Sc (X)| ≥ ω
and the result follows from the previous paragraph and Lemma 3.1 (for i =
2). 
Lemma 4.2. Let X be a space and let V be a cellular family in X. If LX ∩ V 6=
S
∅, then |V | ≤ c(Sc (X)).
Proof. Fix V ∈ V such that V ∩ LX 6= ∅. Then, by Lemma 3.4, Vc+ 6= ∅. Define
U = {h{V, W }ic : W ∈ V }. Thus, U is a cellular family in Sc (X). Since
|V | = |U |, we conclude that |V | ≤ c(Sc (X)). 
Theorem 4.3. Let X be a space. If |LX | ≥ 2, then c(X) ≤ c(Sc (X)).
Proof. It suffices to show that if W is a cellular family in X, then |W | ≤
c(Sc (X)). Note that when W is finite, this inequality is an easy consequence of
Lemma 4.1.
Now, if |W | ≥ ω, fix U, V ∈ τX in such a way that U ∩ LX 6= ∅ = 6 V ∩ LX and
U ∩ V = ∅. Also, set F = {W ∈ W : W ∩ U 6= ∅} and note that the collections
V0 = {W ∩ U : W ∈ F } ∪ {V } and V1 = (W \ F ) ∪ {U }
are cellular families in X with LX ∩ V0 6= ∅ =
6 LX ∩ V1 . Thus, by Lemma 4.2,
S S

|W | = max{|V0 |, |V1 |} ≤ c(Sc (X)).


Theorem 4.4. If X is Tychonoff and abundant in sequences, then
c(Sc (X)) = sup{c(X n ) : n ∈ N}.
Proof. Let Y be a compactification of X. Since Y is compact, [2, Corollary 5.4,
p. 143] and [2, Corollary 5.12, p. 148] imply that c(C L (Y )) = sup{c(Y n ) : n ∈
N}. Now, the fact that X is dense in Y implies that c(Y n ) = c(X n ), for each
n ∈ N, and, on the other hand, Sc (X) is dense in C L (X) so we get the equality
we claimed. 
Corollary 4.5. If (L, ≤) is a connected LOTS having no end-points, then
c(Sc (L)) = sup{c(Ln ) : n ∈ N}.
Proof. The fact that L is Tychonoff follows from [1, Problem 1.7.4(d), p. 57] so
we only need to show that L is abundant in sequences.
First of all, by [1, Problem 6.3.2, p. 373], L is dense in itself, i.e., whenever
a, b ∈ L satisfy a < b, we get (a, b) 6= ∅ (here, (a, b) = {x ∈ L : a < x < b}).
Now, since L has no end-points, for each U ∈ τL \ {∅} there exist p ∈ U
and q, r ∈ L with p ∈ (q, r) ⊆ U . From the previous paragraph we deduce that
6 D. MAYA, P. PELLICER-COVARRUBIAS, AND R. PICHARDO-MENDOZA

there exists {xn : n ∈ ω} ⊆ L in such a way that q < xn < xn+1 < p, for each
n ∈ ω. Hence, according to [1, Problem 6.3.2, p. 373], there is xω ∈ L with
xω = sup{xn : n ∈ ω} and so, {xi : i ≤ ω} is a nontrivial convergent sequence
contained in U . In other words, L is abundant in sequences. 
Let us recall that the Set-theoretic axiom ¬SH (see [5, Definition 4.1, p. 66])
implies, according to [5, (30), p. 90] and [5, Lemma 4.3, p. 66], the existence of L,
a connected LOTS with no end-points, such that c(L) = ω < c(L2 ). Therefore,
it is consistent with the usual axioms for Set Theory that there are Tychonoff
spaces X which are abundant in sequences and satisfy c(X) = ω < c(Sc (X)).
On the other hand, it is also consistent (see [5, Theorem 2.24, p. 61]) that if Y is
a Tychonoff space which is abundant in sequences, then the equality c(Y ) = ω
implies that c(Sc (Y )) = ω.
Question 4.6. Does there exist in ZFC a space X such that c(X) < c(Sc (X))?
Proposition 4.7. If κ is an infinite cardinal and X is the one-point compacti-
fication of the discrete space of size κ, then c(Sc (X)) = ω.
Proof. Let us agree that X is the topological space which results of endowing
the ordinal κ + 1 with the topology having the collection
{X \ F : F ∈ [κ]<ω } ∪ {{ξ} : ξ < κ}
as a base. In particular, Sc (X) = {E ∪ {κ} : E ∈ [κ]ω }.
Now assume that {Uα : α < ω1 } is a family of non-empty open subsets of
Sc (X). We will show that there is B ∈ [ω1 ]ω1 in such a way that {Uα : α ∈ B}
has the finite intersection property (in other words, ω1 is a precaliber for Sc (X)).
Given α < ω1 , use Proposition 3.2 and the fact that Uα 6= ∅ to get Gα ∈ [κ]<ω
and Fα ⊆ Gα satisfying that
∅=
6 Vα = h{X \ Gα } ∪ {{ξ} : ξ ∈ Fα }ic ⊆ Uα .
A straightforward application of [5, Theorem 1.5, p. 49] produces two sets,
B0 ∈ [ω1 ]ω1 and R0 , such that {Gα : α ∈ B0 } is a ∆-system with root R0 ,
i.e., whenever α, β ∈ B0 satisfy α 6= β, we obtain Gα ∩ Gβ = R0 . Use [5,
Theorem 1.5, p. 49] once again to get B1 ∈ [B0 ]ω1 and R1 in such a way that
{Fα : α ∈ B1 } is a ∆-system with root R1 . Observe that if α, β ∈ B1 are such
that α 6= β, then Fα ∩ R0 \ R1 and Fβ ∩ R0 \ R1 are disjoint. Thus, the fact
|R0 | < ω implies that B = {α ∈ B1 : Fα ∩ R0 ⊆ R1S } is uncountable.
Let H ∈ [B]<ω \ {∅}Sbe arbitrary. Fix E ⊆ κ \ {Gξ : ξ ∈ H} with |E| = ω
and set S = {κ} ∪ E ∪ {Fξ : ξ ∈ H}. Suppose that α ∈ H. We will argue that
S ⊆ (X \ Gα ) ∪ Fα , i.e., that if β ∈ H \ {α}, then Fβ \ Fα and Gα are disjoint.
Indeed, since B ⊆ B1 ⊆ B0 , we obtain Fβ \ Fα = Fβ \ R1 and
Gα ∩ Fβ \ Fα = Gα ∩ Gβ ∩ Fβ \ R1 = R0 ∩ Fβ \ R1 = ∅.
T T
In conclusion, S ∈ {Vα : α ∈ H} ⊆ {Uα : α ∈ H}. 
 CARDINAL FUNCTIONS OF Sc (X) 7

Assume Y is the one-point compactification of the discrete space of size κ,

where κ > ω. Then c(Y ) = κ > ω = c(Sc (Y )) and so, we deduce two things:
(1) the assumption |LX | ≥ 2 is essential in Theorem 4.3 and (2) when X is not
abundant in sequences, the equality claimed in Theorem 4.4 may fail.

5. Tightness
The tightness of
S a topological space X, t(X), is the least cardinal κ for which
the equality A = {B : B ∈ [A]≤κ } holds, for all A ⊆ X.
Lemma 5.1. If X is a non-discrete space, then t(X) ≥ ω. In particular, if
Sc (X) 6= ∅, then t(Sc (X)) ≥ ω.
Proof. When x is an accumulation point of X, we get that x ∈ X \ {x} and
{B : B ∈ [X \ {x}]<ω } = X \ {x}. Therefore, t(X) ≥ ω.
S
Finally, assume that Sc (X) 6= ∅. Proposition 3.5 implies that Sc (X) is non-
discrete and so the result follows from the previous paragraph and Lemma 3.1
(for i = 2). 
For a space X, a subset A of X and S ∈ Sc (X), let N(S, A) = {S ∪ {a} : a ∈
A}.
Lemma 5.2. Let X be a space and let A be a nonempty subset of X. If S ∈
Sc (X), then N(S, clX A) ⊆ clSc (X) N(S, A).
Proof. Let z ∈ clX A and U ∈ C(X) be such that S ∪ {z} ∈ hU ic . Choose
U ∈ U satisfying that z ∈ U . The fact that z ∈ clX A guarantees the existence of
x ∈ U ∩A. Then we have that S∪{x} ∈ hU ic ∩N(S, A). This and Proposition 3.2
prove that S ∪ {z} ∈ clSc (X) N(S, A). 
Theorem 5.3. Let X be a nonempty space. Then Sc (X) 6= ∅ if and only if
t(X) ≤ t(Sc (X)).
Proof. Suppose that Sc (X) 6= ∅. Set κ = t(Sc (X)). In order to prove that
t(X) ≤ κ, fix A ⊆ X and x ∈ A. We will show that there exists B ∈ [A]≤κ such
that x ∈ B.
Fix S ∈ Sc (X). We analyze three cases.
Case 1. x 6= lim S.
Set R = S \ {x}, then R ∈ Sc (X). From Lemma 5.2, it follows that R ∪ {x} ∈
N(R, A). Since κ = t(Sc (X)), there exists F ∈ [N(R, A)]≤κ satisfying that
R ∪ {x} ∈ F. Observe that F = N(R, B) for some B ∈ [A]≤κ ; we will prove
that x ∈ B. Let U ∈ τX be such that x ∈ U and fix U1 , U2 ∈ τX such that
x ∈ U1 , R ⊆ U2 , and U1 ∩ U2 = ∅. Note that R ∪ {x} ∈ h{U1 ∩ U, U2 }ic . Since
R ∪ {x} ∈ F, we may take b ∈ B such that R ∪ {b} ∈ F ∩ h{U1 ∩ U, U2 }ic . Thus,
using that R ⊆ U2 , we deduce that b ∈ U1 ∩ U . This shows that x ∈ B.
8 D. MAYA, P. PELLICER-COVARRUBIAS, AND R. PICHARDO-MENDOZA

Case 2. x = lim S and |S ∩ A| = ω.

Set R = S ∩ A. Since x ∈ S ∩ A, we infer that R ∈ Sc (X, x). According
≤κ
S 1, for each r ∈ R \ {x} there exists Br ∈ [A]
to Case such that r ∈ Br . Set
B = {Br : r ∈ R \{x}}. By Lemma 5.1 we know that κ ≥ ω, hence B ∈ [A]≤κ .
Finally, one can easily show that x ∈ B.
Case 3. x = lim S and |S ∩ A| < ω.
Set R = (S \ A) ∪ {x} to get that R ∈ Sc (X, x). Moreover, R ∩ A = {x}.
By Lemma 5.2, R ∈ N(R, A). Since κ = t(Sc (X)), there exists F ∈ [N(R, A)]≤κ
satisfying that R ∈ F. Observe that F = N(R, B) for some B ∈ [A]≤κ . We will
prove that x ∈ B. Let U ∈ τX be such that x ∈ U . Fix V ∈ C(X) with the
following three properties: (a) R ∈ hV i, (b) if V ∈ V is such that x ∈ V , then
V ⊆ U , and (c) if W ∈ V \ {V }, then W ∩ A = ∅. Since R ∈ S F, we may take
b ∈ B satisfying that R ∪ {b} ∈ F ∩ hV ic . In particular, b ∈ V , hence, by
conditions (c) and (b) of our choice of V , we infer that b ∈ V ⊆ U . It follows
that x ∈ B.
The converse is immediate. 
By definition, the character of a space X, χ(X), is the least cardinal κ such
that X has a base of size ≤ κ at x, for each x ∈ X. A straightforward argument
gives t(X) ≤ χ(X).
Corollary 5.4. If X is a space such that Sc (X) 6= ∅, then
t(X) ≤ t(Sc (X)) ≤ χ(X).
Proof. The first inequality follows from Theorem 5.3. For the second inequality,
invoke [6, Theorem 6.6] to get χ(Sc (X)) ≤ χ(X). 
The following example proves that there exists a space X which is abundant
in sequences such that t(X) = t(Sc (X)).
Example 5.5. If X is the long segment (see [1, Problem 3.12.19, p. 237]), then
t(X) = χ(X) = ω1 . Hence, t(Sc (X)) = ω1 .
In our next example we will use the hedgehog of ω1 spines and the sequential
fan of ω1 spines (see Section 2) to produce a space X which is abundant in
sequences and satisfies that t(X) < t(Sc (X)).
Example 5.6. If X is the product J(ω1 ) × 2, where 2 is endowed with the
discrete topology, then t(X) < t(Sc (X)).
Observe that for every B ⊆ X and for every x ∈ clX B there exists a sequence
(xn )n∈ω contained in B whose limit is x (spaces having this property are called
Fréchet-Urysohn); thus t(X) = ω. Now, it is proved in [4, Corollary 1.7 and
Example 1.8, p. 303] that t(S(ω1 )2 ) ≥ ω1 ; moreover the arguments used in the
 CARDINAL FUNCTIONS OF Sc (X) 9

proof of [6, Proposition 3.7] show that S(ω1 )2 embeds into Sc (X). Hence, we
obtain that t(Sc (X)) ≥ ω1 .
Question 5.7. Is there a (normal) space X which is abundant in sequences and
satisfies t(Sc (X)) < χ(Sc (X))?

6. Extent
The extent of a topological space X, denoted by e(X), is the least cardinal κ
for which all closed discrete subspaces of X have cardinality ≤ κ.
Lemma 6.1. Let X be a space. If |X| ≥ ω, then e(X) ≥ ω. In particular, if
Sc (X) 6= ∅, then e(Sc (X)) ≥ ω.
Proof. Fix n < ω and let F ∈ [X]n+1 . Since F is closed and discrete, we deduce
that e(X) > n. Therefore, e(X) ≥ ω.
Finally, assume that Sc (X) 6= ∅. Proposition 3.5 implies that |Sc (X)| ≥ ω
and the result follows from the previous paragraph and Lemma 3.1 (for i =
2). 

Theorem 6.2. Let X be a nonempty space. Then, Sc (X) 6= ∅ if and only if

e(X) ≤ e(Sc (X)).
Proof. For the direct implication it suffices to show that whenever B is a closed
discrete subspace of X, we get that |B| ≤ e(Sc (X)).
Observe that if B is finite, Lemma 6.1 implies easily the above inequality.
So, assume that |B| ≥ ω and fix x ∈ LX . Then, A = B \ {x} is a closed
discrete subspace of X with |A| = |B|. Since X \ A is an open subset of X
containing x, Lemma 3.4 guarantees the existence of S ∈ (X \ A)+ c . Hence,
according to Lemma 3.6, E1 ({A}, S) is a closed subspace of Sc (X) homeo-
morphic to F1 (A). On the other hand, it is a well-known fact that the sub-
space A is homeomorphic to F1 (A); therefore, E1 ({A}, S) is discrete. Thus,
|B| = |E1 ({A}, S)| ≤ e(Sc (X)).
The converse is immediate. 

The following example shows that the inequality in Theorem 6.2 may be strict.
Example 6.3. If S is Sorgenfrey’s line, then
(1) e(S2 ) = e(Sc (S2 )) = c and
(2) e(S) = ω < c ≤ e(Sc (S)).
Indeed, for (1), the fact that {(−x, x) : x ∈ S} is a closed discrete subspace
of S2 implies the following inequalities
 
ω+1 
c ≤ e(S2 ) ≤ e(Sc (S2 )) ≤ Sc (S2 ) ≤  S2

 = c.
10 D. MAYA, P. PELLICER-COVARRUBIAS, AND R. PICHARDO-MENDOZA

For (2), start by recalling that all subspaces of S are separable and so, e(S) =
ω. Finally, the proof of [6, Proposition 3.5] shows that Sc (S) contains a closed
discrete subspace of cardinality c.

7. Dispersion character
The dispersion character of a nonempty topological space X, ∆(X), is the
cardinal number defined by min{|U | : U ∈ τX \ {∅}}.
By the Cantor space we mean the topological product 2ω , where the ordinal
2 is considered as a LOTS.
Theorem 7.1. If X is a space, then any nonempty open subset of Sc (X) con-
tains a copy of the Cantor space. In particular, ∆(Sc (X)) ≥ c.
Proof. According to Proposition 3.2, we just need to fix U ∈ C(X) with hU ic 6=
∅ and embed 2ω into hU ic .
Choose S ∈ hU ic and let {xn : n ≤ ω} be an adequate enumeration of S. Take
U ∈ U satisfying that xω ∈ U . Then, for some m ∈ ω, {xn : n ∈ ω \ (2m)}
Q∞ ⊆ U .
Given n ∈ ω\m, let An be the subspace {x2n , x2n+1 } of X and set A = k=m Ak .
Thus, A is homeomorphic to 2ω and we have that t ∈ A if and only if t is a
function from ω \ m into S in such a way that t(k) ∈ Ak for each k ∈ ω \ m.
Hence, we only need to argue that the function ϕ : A → hU ic given by
ϕ(t) = {xk : k < 2m} ∪ ran(t) ∪ {xω },
for each t ∈ A, is an embedding.
First of all, if s, t ∈ A satisfy that s 6= t, then there is n ∈ ω \ m with
s(n) 6= t(n) and so, s(n) ∈ ϕ(s) \ ϕ(t). In other words, ϕ is one-to-one.
Now, for each n ∈ ω \ m, let πn : A → An be the corresponding projection
map, i. e., πn (t) = t(n) for each t ∈ A. To prove that ϕ is continuous, assume
that V ∈ C(X) and t ∈ A satisfy ϕ(t) ∈ hV i. For each V ∈ V , let us fix an
integer i(V ) with xi(V ) ∈ V ∩ ϕ(t). Now let W ∈ V be such that xω ∈ W
and assume that ` ∈ ω satisfies that {m} ∪ {i(V ) : V ∈ V } ⊆ ` and {xk : k ∈
ω \ (2`)} ⊆ W . Hence,
\
G = {πn −1 [{t(k)}] : k ∈ ` \ m}
is an open subset of A with t ∈ G. Moreover, if s ∈ G, then s[` \ m] = t[` \ m]
and so, xi(V ) ∈ V ∩ ϕ(s), whenever V ∈ V ; on the
S other hand, the assumption
ϕ(t) ∈ hV ic gives {xk S: k < 2m} ∪ s[` \ m] ⊆ V and so, our choice for `
guarantees that ϕ(s) ⊆ V . Therefore, ϕ[G] ⊆ hV ic .
Finally, note that A is a compact space and ϕ[A] is Hausdorff, to conclude
that ϕ : A → ϕ[A] is closed. 
Lemma 7.2. Let X be a space and let U ∈ C(X) be such that hU ic 6= ∅. Then
|hU ic | = max ({|U | : U ∈ U } ∪ {|Sc (U )| : U ∈ U }) (7.1)
 CARDINAL FUNCTIONS OF Sc (X) 11

Proof. For each U ∈ U define QU = {S ∈ hU ic : lim S ∈ U }. Fix U ∈ U

such that QU 6= ∅. Then, Sc (U ) 6= ∅; hence, by Proposition 3.5, |Sc (U
S )| ≥ ω.
Thus, the equivalence, S ∈ QU if and only if S ∩ U ∈ Sc (U ), S ⊆ U and
S ∩ V ∈ [V ]<ω \ {∅}, for every V ∈ U \ {U }, gives

|QU | = |Sc (U )|·max{|[V ]<ω | : V ∈ U \{U }} = |Sc (U )|·max{|V | : V ∈ U \{U }}.

Therefore, since {QU : U ∈ U } \ {∅} is a partition of hU ic , we conclude that

the equality in (7.1) holds. 

Corollary 7.3. Let X be a space and let U ∈ C(X) be such that hU ic 6= ∅.

Then
max{|U | : U ∈ U } ≤ |hU ic | ≤ max{|U |ω : U ∈ U }.

Corollary 7.4. Let X be a nonempty space. Then Sc (X) 6= ∅ if and only if

∆(X) ≤ ∆(Sc (X)).

Proof. Assume first that Sc (X) 6= ∅. If V ∈ C(X) is such that hV ic 6= ∅ and

V ∈ V is arbitrary, then Corollary 7.3 implies that ∆(X) ≤ |V | ≤ |hV ic |. Thus,
the inequality ∆(X) ≤ ∆(Sc (X)) follows from Proposition 3.2.
The converse is immediate. 

Example 7.5. Let 0 < α ≤ ω and let U ∈ C(Rα ) be such that hU ic 6= ∅.

Since |U | = c = |U |ω for each U ∈ U , Corollary 7.3 implies that |hU ic | = c.
Therefore, by Proposition 3.2, ∆(Sc (Rα )) = c = ∆(Rα ).

Note that if Sc (X) 6= ∅ and X is not crowded, then its dispersion character
is strictly less than that of Sc (X) by Theorem 7.1.

Examples of crowded spaces X such that ∆(X) < ∆(Sc (X)) can be produced
as follows.

Example 7.6. Consider the ordinal ω with the discrete topology and denote
by ω ∗ the remainder of the Stone-Čech compactification of ω. It follows from
[1, theorems 3.6.13 and 3.6.14, p. 175] that ∆(ω ∗ ) = 2c . Now, given a cardinal
κ > 2c , let κ 2 be the topological product of κ copies of the discrete space of
size 2 and note that 2κ = |κ 2|. Set X = ω ∗ ⊕ κ 2 to get ∆(X) = ∆(ω ∗ ) = 2c
and fix U ∈ C(X) such that hU ic 6= ∅. Since Sc (ω ∗ ) = ∅, we infer that
U ∩ κ 2 6= ∅ for some U ∈ U . Hence, 2κ = |U | ≤ |hU ic | (see Corollary 7.3) and
so, ∆(X) < 2κ ≤ ∆(Sc (X)).

Example 7.7. If Q denotes the set of all rational numbers, then Q is crowded
and ∆(Q) = ω < c ≤ ∆(Sc (Q)) (see Theorem 7.1).
12 D. MAYA, P. PELLICER-COVARRUBIAS, AND R. PICHARDO-MENDOZA

8. Weighing Sc (X)
A family N of subsets of a space X is a network for X if τX ⊆ { E : E ⊆
S
N }. Hence, the net weight of X is, by definition, the least cardinality of a
network for X and it will be denoted by nw(X).
Lemma 8.1. If X is a space such that |X| ≥ ω, then nw(X) ≥ ω. In particular,
if Sc (X) 6= ∅, then nw(Sc (X)) ≥ ω.
Proof. The result follows from the fact that c(Y ) ≤ nw(Y ) for each space Y and
Lemma 4.1. 
The proof of our next result is straightforward.
Proposition 8.2. Let X be a space. If Y ⊆ X, then nw(Y ) ≤ nw(X).
Theorem 8.3. If X is a space such that Sc (X) 6= ∅, then nw(X) ≤ nw(Sc (X)).
Proof. Choose S ∈ Sc (X) in such a way that |X \S| ≥ ω. Next fix a network N
for X \ S. Note that |N | ≥ ω (Lemma 8.1). Set N1 = N ∪ {{x} : x ∈ S}, then
N1 is a network for X and |N1 | = |N |. This implies that nw(X) ≤ nw(X \ S).
Moreover, [3, Lemma 1.1, p. 796] guarantees that X \ S embeds into Sc (X).
Therefore, Proposition 8.2 gives nw(X) ≤ nw(X \ S) ≤ nw(Sc (X)). 
Example 8.4. A space X such that nw(X) < nw(Sc (X)).
In the plane define a local base for each (a, b) as follows: if b 6= 0, then (a, b)
will have its usual local base. Moreover, for each a ∈ R and each r > 0 set
Ma,r = {(a, 0)} ∪ {(x, y) ∈ R2 : |x − a| < r ∧ |y| < |x − a|}.
For each point of the form (a, 0) take the local base given by {Ma,r : r > 0}. Let
X be the topological space obtained by endowing the plane with the resulting
topology.
Set Z1 = {(x, 0) : x ∈ R} and Z2 = X \ Z1 . Observe that the induced
topologies on both Z1 and Z2 coincide with the topologies induced by the usual
topology of the plane. For each j ∈ {1, 2} fix a countable base Bj for Zj , then
B1 ∪ B2 is a countable network for X. In other words, nw(X) = ω.
Next, for each a ∈ R set
Sa = {(a, 0)} ∪ (a − 21n , 2n+1
1
) : n ∈ ω ∪ (a + 21n , 2n+1
1
 
):n∈ω .
Note that Sa ∈ Sc (X) and define G = {Sa : a ∈ R}.
The claim below and Proposition 8.2 imply that c = nw(G ) ≤ nw(Sc (X)).
Claim. G is discrete.
Indeed, fix a ∈ R and set U = Ma, 23 , it follows that {Sa } ⊆ Uc+ ∩ G . In order
to prove that Uc+ ∩ G ⊆ {Sa } let t ∈ R be such that St ∈ Uc+ ∩ G . Note that
t ∈ (a − 32 , a + 32 ).
 CARDINAL FUNCTIONS OF Sc (X) 13

We will argue that S neither t < a1 nor t > a. Suppose that t < a, then
t ∈ (a − 23 , a − 1] ∪ n∈ω [a − 21n , a − 2n+1 ]. If t ∈ (a − 23 , a − 1], a straightforward
calculation gives |t + 1 − a| < 12 , which implies that (t + 1, 12 ) ∈ St \ U . Now, if
t ∈ [a− 21n , a− 2n+1
1
] for some n ∈ ω, an analogous argument gives (t+ 21n , 2n+1 1
)∈
St \ U . This contradiction shows that t 6< a however, arguing similarly one can
show that t 6> a. Hence, t = a. Therefore St = Sa and the claim is proved.
Given X, a space, and S, T ∈ Sc (X) we will say that S is almost contained
in T (in symbols, S ⊆∗ T ) if |S \ T | < ω. Thus, a family C ⊆ Sc (X) is cofinal
in Sc (X) if for each S ∈ Sc (X) there is T ∈ C with S ⊆∗ T . The minimum size
of a cofinal subset of Sc (X) is denoted by cf c (X). Clearly, cf c (X) ≤ |Sc (X)|
for any space X.
Proposition 8.5. If X is a space, then |LX | ≤ cf c (X).
Proof. Assume that C is a cofinal subset of Sc (X) of minimum size. For each
y ∈ LX fix Sy ∈ Sc (X, y) and Ty ∈ C such that Sy ⊆∗ Ty . Observe that Ty → y.
Hence, if ϕ : LX → C is given by ϕ(y) = Ty , then ϕ is one to one. 
Example 8.6. Let κ be an infinite cardinal and recall that S(κ) denotes the
sequential fan of κ spines (see Section 2). We claim that cf c (S(κ)) = κ.
Let q : (ω + 1) × κ → S(κ) be the natural quotient map and let z be the only
non-isolated point of S(κ).
Assume that E ⊆ Sc (S(κ)) satisfies |E | < κ and, for each T ∈ E , set
T † = {α < κ : (T \ {z}) ∩ q[(ω + 1) × {α}] 6= ∅}
to obtain a finite subset of κ. Hence, the set {T † : T ∈ E } has fewer than κ
S
many members and so, there is δ ∈ κ in such aSway that z is the only point that
the convergent sequence q[(ω + 1) × {δ}] and E have in common. Thus, E is
not cofinal in Sc (S(κ)) and so, κ ≤ cf c (S(κ)).
The remaining inequality follows from the fact that
{q[(ω + 1) × F ] : F ∈ [κ]<ω \ {∅}}
is a cofinal subset of Sc (S(κ)).
An immediate consequence of our previous example is that cf c (X) may be
any infinite cardinal, even singular. We also obtain that cf c (S(ω)) = ω < c =
|Sc (S(ω))|.
Proposition 8.7. For a topological space X,
nw(Sc (X)) ≤ min{nw(X)ω , cf c (X) + nw(X)}.
Proof. Fix N , a network for X of cardinality nw(X).
To prove that nw(Sc (X)) ≤ nw(X)ω we will argue that {hE ic : E ∈ [N ]≤ω }
is a network for Sc (X). Assume that U ∈ C(X) and S ∈ hU ic are arbitrary. For
14 D. MAYA, P. PELLICER-COVARRUBIAS, AND R. PICHARDO-MENDOZA

each U ∈ U and x ∈ S ∩ U choose ExU ∈ N in such a way that x ∈ ExU ⊆ U .

Hence, E = {ExU : U ∈ U ∧ x ∈ S ∩ U } is a countable subset of N and
S ∈ hE ic ⊆ hU ic .
For the other inequality, let us assume that C is a cofinal subset of Sc (X) of
minimum size. We will show that
N = {h{T \ F } ∪ E ic : T ∈ C ∧ F ∈ [T ]<ω ∧ E ∈ [N ]<ω }
is a network for Sc (X).
Assume that U ∈ C(X) and S ∈ hU ic are arbitrary. Fix W ∈ U and T ∈ C
in such a way that a = lim S ∈ W and S ⊆∗ T . Thus, lim T = a and so, the sets
F = T \ W and H = S \ (T ∩ W ) are finite. Now, for each x ∈ H let Ux ∈ U and
Nx ∈ N be so that x ∈ Nx ⊆ Ux . Therefore, N = h{T \ F } ∪ {Nx : x ∈ H}ic is
a member of N satisfying S ∈ N ⊆ hU ic . 

In connection with Proposition 8.7; if Z is a second countable space such that

|LZ | > ω (e. g. a sphere), it follows by Lemma 3.3 and Proposition 8.5 that
nw(Sc (Z)) = ω < min{nw(Z)ω , cf c (Z) + nw(Z)}.
Proposition 8.8. Let X be a locally compact space, a LOTS or a metric space.
If Sc (X) 6= ∅, then nw(X) = nw(Sc (X)).
Proof. It is known that w(X) = nw(X) (see [1, Theorem 3.3.5, p. 149], [1,
Problem 3.12.4(d),p. 222], and [1, Theorem 4.1.15, p. 255]).Thus, Lemma 3.3
gives that nw(Sc (X)) ≤ w(Sc (X)) = w(X) = nw(X) and the result follows
from Theorem 8.3. 

Assume τ and σ are topologies on a given set E. Then, the symbol (E, σ) ≤
(E, τ ) will be used whenever σ ⊆ τ . Thus, if X is a Tychonoff space, we define the
i-weight of X to be the least cardinal of the form w(Y ), where Y is a Tychonoff
space satisfying Y ≤ X. We will employ iw(X) to denote this cardinal.
Note that according to Lemma 3.1, if X is a Tychonoff space, then Sc (X) is
a Tychonoff space as well.
Proposition 8.9. If X is a Tychonoff space, then iw(Sc (X)) ≤ iw(X).
Proof. When Sc (X) = ∅, we get iw(Sc (X)) = 0 and so, the inequality holds
trivially. Let us assume that Sc (X) 6= ∅.
Assume Y is a Tychonoff space satisfying iw(X) = w(Y ) and Y ≤ X. Hence,
all convergent sequences in X are convergent sequences in Y so, denote by E
the topological space which results of endowing the set Sc (X) with the relative
topology from Sc (Y ). Since Y is Tychonoff, E is Tychonoff as well and, more-
over, if G is a Vietoris open set in Sc (Y ), then G ∩ E is a Vietoris open set in
Sc (X), i.e., E ≤ Sc (X).
 CARDINAL FUNCTIONS OF Sc (X) 15

From the previous paragraph we deduce that iw(Sc (X)) ≤ w(E). To complete
our argument, invoke Lemma 3.3 to obtain w(E) ≤ w(Sc (Y )) = w(Y ) = iw(X).

Question 8.10. Is there a Tychonoff space X such that iw(Sc (X)) < iw(X)?
Assume that P is a family of nonempty open subsets of the topological space
X. P will be called a π-base for X if each nonempty open subset of X contains
a member of P. Thus, the π-weight of X, πw(X), is the least cardinality of a
π-base for X. On the other hand, given z ∈ X, we will say that P is a π-base
for X at z if whenever z ∈ G ∈ τX , there is P ∈ P with P ⊆ G. Hence, the
π-character of X at z, πχ(X, z), is the least cardinality of a π-base for X at z.
Finally, define πχ(X) = sup{πχ(X, y) : y ∈ X}.
Lemma 8.11. Let X be a space. If |X| ≥ ω, then πw(X) ≥ ω. In particular,
if Sc (X) 6= ∅, then πw(Sc (X)) ≥ ω.
Proof. This is an immediate consequence of the inequality c(X) ≤ πw(X) and
Lemma 4.1. 
Let us agree that if A is a family of subsets of some set X and Y ⊆ X, then
the trace of A over Y is A  Y = {Y ∩ A : A ∈ A }.
Now, in order to obtain a formula relating the π-weight of the hyperspace
of convergent sequences with the π-weight of the ground space, we need to
introduce the following notion.
Given a topological space X, a nonempty subspace Y ⊆ X, and P ⊆ τX , we
will say that P is a π-base for Y modulo X if ∅ ∈ / P  Y and for each G ∈ τX
satisfying G ∩ Y 6= ∅ there is P ∈ P with P ⊆ G.
For a nonempty subspace Y of a topological space X, the least cardinality of
a π-base for Y modulo X is denoted by πwX (Y ). Since the trace over Y of any
π-base for Y modulo X is a π-base for Y , we get πw(Y ) ≤ πwX (Y ).
Example 8.12. Observe that if p ∈ X, then being a π-base for {p} modulo X
is equivalent to being a local base for X at p. In particular, if p is a non-isolated
point of a space X, then
πw({p}) = 1 < πwX ({p}).
Proposition 8.13. If X has a nontrivial convergent sequence, then
πw(Sc (X)) = πw(X) + πwX (LX ).
Proof. Notice that our hypothesis on X and Lemma 8.11 imply that the cardinals
πw(X) and κ = πw(Sc (X)) are infinite.
Apply Proposition 3.2 to get {Uα : α < κ} ⊆ C(X) in such a way that
{hUα ic : α < κ} is a π-base for Sc (X).
Claim 1. α<κ Uα is a π-base for X.
S
16 D. MAYA, P. PELLICER-COVARRUBIAS, AND R. PICHARDO-MENDOZA

Let G be a nonemtpy open subset of X. According to Lemma S 3.4, if G∩LX 6=

∅, then G+ c 6
= ∅ and so, for some α < κ, hUα i c ⊆ G+
c ; thus, Uα ⊆ G (see
[6, Lemma 6.3]). Now, when G ∩ LX = ∅, fix S ∈ Sc (X) and z ∈ G. Clearly,
z 6= lim S and so there are disjoint U, V ∈ τX in such a way that z ∈ U and
S \ {z} ⊆ V . Set O = h{U ∩ G, V }ic , note that S ∪ {z} ∈ O, and let α < κ
and T be so that T ∈ hUα ic ⊆ O. Hence, lim T ∈ (U ∩ G) ∪ V , but given our
assumption on G, we get that lim T ∈ V ; in particular, T ∩ (U ∩ G) is finite and
so, by [6, Lemma 6.2], there is W ∈ Uα with W ⊆ G ∩ U ⊆ G.
Claim 2. {P ∈ α<κ Uα : P ∩ LX 6= ∅} is a π-base for LX modulo X.
S

Indeed, if G ∈ τX has nonempty intersection with LX , then G+c 6= ∅ (see

Lemma 3.4). Hence, by [6, Lemma 6.3], Uα ⊆ G for some α < κ, and given
S
that hUα ic 6= ∅, there is P ∈ Uα with P ∩ LX 6= ∅.
From these two claims we obtain πw(X) + πwX (LX ) ≤ κ. For the remaining
inequality we will argue that if P and P † are a π-base for X and a π-base for
LX modulo X, respectively, then
P = {hU ic : U ∈ [P ∪ P † ]<ω ∧ U ∩ P † 6= ∅}
is a π-base for Sc (X). First, Lemma 3.4 implies that P is a family of nonempty
open subsets of Sc (X).
Let V ∈ C(X) be such that hV ic 6= ∅. Fix G ∈ V and G∗ ∈ P † with
G ∩ LX 6= ∅ and G∗ ⊆ G. Also, for each V ∈ V \ {G} let V ∗ ∈ P be so that
V ∗ ⊆ V . Hence, h{V ∗ : V ∈ V }ic is a member of P contained in hV ic . 
Note that if Y is a dense subspace of the topological space X, then any π-base
for X is a π-base for Y modulo X and so, πwX (Y ) ≤ πw(X). As a consequence
of this remark and Proposition 8.13 we obtain the following result.
Corollary 8.14. If X is abundant in sequences, then πw(Sc (X)) = πw(X).
Example 8.15. Let X be the sequential fan of ω spines S(ω). If p is the only
non-isolated point of X, then {{x} : x ∈ X \ {p}} is a π-base for X; hence
πw(X) = ω. On the other hand, in Example 8.12, we argued that local π-bases
for LX modulo X coincide with local bases for X at p; thus, a reasoning similar
to the one exposed in [1, Example 1.4.17, p. 33] gives us πwX (LX ) > ω and so,
πw(X) < πwX (LX ).
Example 8.16. Consider ω + 1 as a LOTS and ω1 \ (ω + 1) as a discrete space.
Now set X = (ω + 1) ⊕ (ω1 \ (ω + 1)) and observe that LX = {ω}. Since
{(ω + 1) \ n : n ∈ ω} is a local base for X at ω, we deduce that πwX (LX ) = ω.
Given that the collection {{α} : α ∈ ω1 \ (ω + 1)} is contained in any π-base for
X, we conclude that
πw(X) > πwX (LX ).
 CARDINAL FUNCTIONS OF Sc (X) 17

Question 8.17. Determine necessary and sufficient conditions on a space X in

order to satisfy the inequality πw(X) ≤ πwX (LX ) (respectively, πw(X) ≥
πwX (LX )).
Now we turn our attention to the cardinal function πχ(X).
Lemma 8.18. If z is not an isolated point of a space X, then πχ(X, z) ≥ ω.
Proof. Let us note that if P is a finite family of nonempty open subsets of X,
then, for each P ∈ P, there is xP ∈ P \ {z} and so, X \ {xP : P ∈ P} is
an open neighborhood of z which contains no member of P. This shows that
πχ(X, z) ≥ ω. 
Theorem 8.19. If X is a space such that Sc (X) 6= ∅, then πχ(X) ≤ πχ(Sc (X)).
Proof. Set κ = πχ(Sc (X)), then Proposition 3.5 and Lemma 8.18 give κ ≥ ω.
Fix z ∈ X. In order to prove that πχ(X, z) ≤ κ, let S0 ∈ Sc (X) and set
S = S0 ∪ {z}. We take two cases.
Case 1. z = lim S.
For each F ∈ [S \ {z}]<ω , let EF ⊆ C(X) be so that |EF | ≤ κ and {hU ic :
U ∈ EF } is a π-base for Sc (X) at S \ F . Observe that the collection
[n[ o
EF : F ∈ [S \ {z}]<ω
has at most κ members so, to complete the proof of this case, we only need to
argue that it is a π-base for X at z.
Assume that z ∈ G ∈ τX and let F = S \ G. Then, S \ F ∈ G+ c and so there
exists U ∈ EF with ∅ = 6 hU ic ⊆ G+c . Hence, for each P ∈ U , by [6, Lemma 6.3]
we obtain that P ⊆ G.
Case 2. z 6= lim S.
Let E ⊆ C(X) be so that |E| ≤ κ and {hU ic : U ∈ E} is a π-base for Sc (X)
at S.
Fix two disjoint open sets, W and V , such that z ∈ W and S \ {z} ⊆ V .
Define n [ o
G = G∩W : G∈ E ∧ G ∩ W 6= ∅ .
Note that |G | ≤ κ. We will prove that G is a π-base for X at z. Assume that
z ∈ U ∈ τX . Since S ∈ h{U ∩ W, V }ic , there exists U ∈ E satisfying that
∅ 6= hU ic ⊆ h{U ∩ W, V }ic . If T ∈ShU ic , we may take G ∈ U such that
T ∩ (U ∩ W ) ∩ G 6= ∅. Given that G ⊆ U ⊆ (U ∩ W ) ∪ V (see [6, Lemma 6.3]),
we conclude that G ∩ W ⊆ U ∩ W ⊆ U . 
Proposition 8.20. Let X be the sequential fan of ω spines. Then, πχ(X) =
ω < πχ(Sc (X)).
18 D. MAYA, P. PELLICER-COVARRUBIAS, AND R. PICHARDO-MENDOZA

Proof. Let p be the only accumulation point of X. Since {{x} : x ∈ X \ {p}}

is a π-base for X at p, Lemma 8.18 gives πχ(X) = ω. Set κ = πχ(Sc (X)). By
Proposition 3.5 and Lemma 8.18 we know that κ ≥ ω, we will prove that κ > ω.
Fix S ∈ Sc (X), then p ∈ S. Let E ⊆ C(X) be so that |E| ≤ κ and {hU ic :
U ∈ E} is a π-base for Sc (X) at S. For each U ∈ E, since hU ic 6= ∅, we may
take UU ∈ U with p ∈ UU . Observe that the collection
G = {UU \ F : (U , F ) ∈ E × [S \ {p}]<ω }
has at most κ members.
We will prove that G is a local base for X at p. Assume that p ∈ V ∈ τX
and define V = {V } ∪ {{x} : x ∈ S \ V }. Since S ∈ hV ic , there exists W ∈ E
satisfying that hW ic ⊆ hV ic . It remains to prove that p ∈ UW \ (S \ V ) ⊆ V .
Clearly p ∈ UW \ (S \ V ). Next, let z ∈ UWS\ (S \ V ) and fix T ∈ hW ic , then
T ∪ {z} ∈ hW ic ⊆ hV ic , in particular z ∈ V . By our assumption on z we
deduce that z ∈ V .
Finally, since X is not first countable at p (see [1, Example 1.4.17, p. 33]), we
conclude that ω < |G | ≤ κ. 
Theorem 8.21. Let X be a space. If LX ⊆ intX (clX LX ), then πχ(Sc (X)) ≤
πχ(X).
Proof. Suppose T ∈ Sc (X) is arbitrary and, for each x ∈ T , let Px be a π-base
for X at x satisfying that πχ(X, x) = |Px |. Set a = lim T . Given G ∈ τX
with a ∈ G, since a ∈ G ∩ intX (clX LX ), there exists P ∈ Pa satisfying that
P ⊆ G ∩ intX (clX LX ) ⊆ G. Thus, without loss of generality, S we may assume
that each element P of Pa is a subset of clX LX . Define P = {Px : x ∈ T }.
We will prove that P = {hU ic : U ∈ [P]<ω ∧ U ∩ Pa 6= ∅} is a π-base for
Sc (X) at T . Assume that U is finite subset of P and that U ∈ U ∩Pa . By our
assumption on Pa we have that U ∩LX 6= ∅. It follows from this and Lemma 3.4
that all members of P are nonempty open subsets of Sc (X). Now suppose that
W ∈ V ∈ C(X) satisfy that T ∈ hV ic and a ∈ W . Then, there exists W ∗ ∈ Pa
with W ∗ ⊆ W . Let e be a choice function for {T ∩ V : V ∈ V \ {W }} and,
for each V ∈ V \ {W }, fix V ∗ ∈ Pe(T ∩V ) in such a way that V ∗ ⊆ V . Hence,
h{V ∗ : V ∈ V }ic is a member of P which is contained in hV ic .
Finally, since ω ≤ |P| ≤ πχ(X) (Lemma 8.18), we conclude that πχ(Sc (X)) ≤
πχ(X). 
The proof of our next result follows from theorems 8.19 and 8.21.
Corollary 8.22. Let X be a space. If X is abundant in sequences, then πχ(X) =
πχ(Sc (X)).
In [6, Question 6.9] the authors ask if the equality χ(X) = χ(Sc (X)) holds
for any space X having a non-trivial convergent sequence. Actually, in view of
 CARDINAL FUNCTIONS OF Sc (X) 19

[6, Theorem 6.6] and the proof of [6, Corollary 6.8], a positive answer to this
question follows from our result below.
Proposition 8.23. Let X be a space with LX = {z}, then χ(X, z) ≤ χ(Sc (X)).
Proof. Our argument is a slight modification of the one used in the proof of
Theorem 8.19.
Set κ = χ(Sc (X)) and fix S ∈ Sc (X). Thus, S → z and so, for each
F ∈ [S \ {z}]<ω there is EF ∈ [C(X)]≤κ in such a way that {hU ic : U ∈ EF } is
a base for Sc (X) at S \ F . Hence,
n[ o
U : ∃F ∈ [S \ {z}]<ω (U ∈ EF )
is a family of at most κ many open subsets of X, each one of them containing
z. We claim that this collection is a base for X at z. Indeed, if z ∈ G ∈ τX ,
then F = S \ G is a finite set and S \ F ∈ G+c ; from here we get the
S existence of
U ∈ EF satisfying hU ic ⊆ G+ c . Therefore (see [6, Lemma 6.3]), U ⊆ G. 

9. Pseudocharacter
Given a space X and aTpoint z ∈ X, we say that a collection P ⊆ τX is a
pseudobase for X at z if P = {z}. The symbol ψ(X, z) will represent the
pseudocharacter of X at z, i.e., the least cardinality of a pseudobase for X at z.
Also, the pseudocharacter of X, ψ(X), is the cardinal sup{ψ(X, x) : x ∈ X}.
Lemma 9.1. If z is a non-isolated point of a topological space X, then ψ(X, z) ≥
ω. In particular, if Sc (X) 6= ∅, then ψ(Sc (X)) ≥ ω.
Proof. Observe that if a pseudobase P of X at z were finite, then {z} would be
an isolated point of X, a contradiction. Therefore, ψ(X, z) ≥ ω.
Finally, assume that Sc (X) 6= ∅. Proposition 3.5 implies that each point of
Sc (X) is non-isolated and the result follows from the previous paragraph and
Lemma 3.1 (for i = 2). 
Proposition 9.2. If X is a space with a nontrivial convergent sequence, then
ψ(X) = ψ(Sc (X)).
Proof. First we will show that ψ(Sc (X)) ≤ ψ(X). Set κ = ψ(X) and fix
S ∈ Sc (X). Assume that lim S = a and, for each x ∈ S, let Px be a pseudobase
for X at x with |Px | ≤ κ. Note that we may, and shall, assume that
X \ (S \ {x}) ∈ Px , whenever x ∈ S \ {a}. (9.1)
Now define P = {Px : x ∈ S} and P = {hU ic : U ∈ [P]<ω ∧ S ∈ hU ic }.
S
Given that κ ≥ ω (see Lemma 9.1) we obtain that |P| ≤ κ. Thus, we only need
P is a pseudobase for Sc (X) at S.
to argue that T
Since S ∈ P, let T ∈ Sc (X) \ {S} be arbitrary and consider two cases.
First, when there is z ∈ T \S, we fix Ga ∈ Pa with z ∈
/ Ga and for each x ∈ S\Ga
20 D. MAYA, P. PELLICER-COVARRUBIAS, AND R. PICHARDO-MENDOZA

we choose Gx ∈ Px such that z ∈ / Gx ; hence, U = {Gx : x ∈ {a} ∪ (S \ Ga )} is

a member of [P]<ω satisfying that S ∈ hU ic and T ∈ / hU ic .
Now, when T ⊆ S and there is y ∈ S \ T , we deduce that y 6= a and so (recall
(9.1)), there exists Gy ∈ Py in such a way that Gy ∩ T = ∅. Fix arbitrary
Ga ∈ Pa and Gx ∈ Px , for each x ∈ S \ (Ga ∪ {y}). Then, V = {Gx : x ∈
(S \ Ga ) ∪ {a, y}} satisfies hV ic ∈ P and T ∈ / hV ic .
In order to prove that ψ(X) ≤ ψ(Sc (X)), set λ = ψ(Sc (X)). It follows from
Proposition 3.5 and Lemma 9.1 that λ ≥ ω. We will argue that if z ∈ X, then
ψ(X, z) ≤ λ.
Suppose S ∈ Sc (X) satisfies z ∈ S and let E ⊆ C(X) be so thatS|E| ≤ λ and
{hU ic : U ∈ E} is a pseudobase T for Sc (X) at S. Then, P = {U ∈T E : z ∈ U }
satisfies P ∈ [τX ]≤λ and z ∈ P, so we only need to prove that P ⊆ {z}.
Fix y ∈ X \ {z}. We analyze two cases. If y ∈ / S, then T = S ∪ {y} ∈
Sc (X) \ {S} and so there exists U ∈ E with T ∈ / hU ic . GivenSthat S intersects
each element of U (recall that S ∈ hU ic ), we deduce that y ∈ / U . Fix U ∈ U
satisfying that z ∈ U to get that U ∈ P and y ∈ / U.
If y ∈ S, fix t ∈ {z, y} \ {lim S}, then R = S \ {t} S ∈ Sc (X) \ {S}, and so
there exists U ∈ E with R ∈ / hU ic . Since R ⊆ S ⊆ U , there is W ∈ U such
that W ∩ R = ∅. Moreover, the fact that S ∈ hU ic implies that S ∩ W 6= ∅;
hence, W ∩ S = {t}. In the case that t = z, we have that W ∈ P and y ∈ / W;
if t = y, let V ∈ U be such that z ∈ V . Then V ∈ P and, since U is cellular,
we conclude that {y} = W ∩ S ⊆ X \ V . 

10. Spaces of continuous functions and Lindelöf degree

Assume X and Y are topological spaces. As usual, we will denote by C(X, Y )
the set of all continuous functions from X into Y . Also, given K, a compact
subset of X, and U , an open subset of Y , we define the set [K, U ] by the formula
f ∈ [K, U ] ↔ f ∈ C(X, Y ) ∧ f [K] ⊆ U.
We will use the symbol Ck (X, Y ) to represent the topological space which results
of endowing C(X, Y ) with the topology which has the collection
{[K, U ] : K ∈ K (X) ∧ U ∈ τY }
as a subbase (this topology is called the compact-open topology on C(X, Y )).
For the following results we will consider the ordinal ω + 1 endowed with the
order topology. The space Ck (ω + 1, X) will be denoted by C(X).e Let Ek (X)
be the subspace of C(X)
e consisting of all mappings whose range is infinite.
Let X be a space and define ` : Sc (X) → X by `(S) = lim S.
Proposition 10.1. Assume that X is a space. Define ρ : Ek (X) → Sc (X) by
ρ(f ) = ran(f ). Then,
(1) ρ is a continuous surjection and
 CARDINAL FUNCTIONS OF Sc (X) 21

(2) ` ◦ ρ is continuous; thus, ρ is not necessarily a quotient map.

Proof. It is immediate that ρ is onto. Now, in order to prove that ρ is contin-
uous, suppose that f ∈ Ek (X) and U ∈ C(X) satisfy that ρ(f ) ∈ hU ic (see
Proposition 3.2). Since ρ(f ) → f (ω), we deduce that f −1 [U ] is a compact subset
of ω + 1, for each U ∈ U ; therefore,
\
V = Ek (X) ∩ {[f −1 [U ], U ] : U ∈ U }
is an open subset of Ek (X) which contains f and satisfies that ρ[V ] ⊆ hU ic .
Let us argue the continuity of `◦ρ. Observe that if f ∈ Ek (X) and (`◦ρ)(f ) ∈
U ∈ τX , then, by letting W = Ek (X)∩[{ω}, U ], we get that W is an open subset
of Ek (X) such that f ∈ W and (`◦ρ)[W ] ⊆ U . Now, in order to argue the second
part of (2), we are going to prove that ` is continuous if and only if ` is constant.
Assume ` is not constant and let S, T ∈ Sc (X) be so that `(S) 6= `(T ). Fix
adequate enumerations, {xi : i ≤ ω} and {yi : i ≤ ω}, of S and of T , respectively.
Thus, xω 6= yω . Now observe that, for each n ∈ ω, the set
Qn = {yi : i ∈ (ω + 1) \ n} ∪ {xi : i ∈ n}
satisfies that Qn ∈ Sc (X, yω ); in particular, `(Qn ) = yω .
On the other hand, the sequence (Qn )n∈ω converges to Q = S ∪ {yω } in
Sc (X) and since `(Q) = xω , this proves that ` is not continuous. 
Proposition 10.2. If X is a space such that Sc (X) 6= ∅, then Ek (X) is not a
closed subset of C(X).
e

Proof. Fix S ∈ Sc (X) and let a = lim S. We claim that f = (ω + 1) × {a}

belongs to the closure of Ek (X). Indeed, if for some m ∈Tω we have {Ki :
i ≤ m} ⊆ K (ω + 1) and {Ui : i ≤ m} ⊆ τX with f ∈ i≤m [Ki , Ui ], then
T T
a ∈ i≤m Ui and so, by fixing a bijection g from ω + 1 onto S ∩ i≤m Ui with
g(ω) = a, we get
\
g ∈ Ek (X) ∩ [Ki , Ui ].
i≤m

Given a space X, let us denote the topological product (ω + 1) × X by X. e
Hence, each member of C(ω + 1, X) is a countable subset of X (we are adopting
e
the set-theoretic convention of identifying a function with its graph); moreover,
if f ∈ C(ω + 1, X), then the map from ω + 1 into X e given by i 7→ (i, f (i)) is
continuous and so f converges to (ω, f (ω)) in X.e In particular, C(ω + 1, X) ⊆
Sc (X).
e So, a natural question is, what is the topology that C(ω + 1, X) inherits
from Sc (X)?
e

Theorem 10.3. For any space X, we have the following:

22 D. MAYA, P. PELLICER-COVARRUBIAS, AND R. PICHARDO-MENDOZA

(1) C(ω + 1, X) is a nowhere dense subset of Sc (X)

e and its closure is the
set

F = {S ∈ Sc (X)
e : ∀n ∈ ω (|S ∩ ({n} × X) | = 1)}.

(2) The relative topology of C(ω +1, X) in Sc (X)

e is the compact-open topol-
ogy.

Proof. To simplify our notation, set Y = C(ω + 1, X) and, for each i ≤ ω, let
Vi = {i} × X.
We will argue first that F is closed: if S ∈ Sc (X)
e \ F, there is m < ω with
|S ∩ Vm | 6= 1. When S ∩ Vm = ∅, we obtain that X e \ Vm is an open subset of
Xe satisfying S ∈ (Xe \ Vm ) ⊆ Sc (X)
+ e \ F. In case that |S ∩ Vm | ≥ 2, there
c
exist U and W , disjoint open subsets of X,e with U ∪ W ⊆ Vm , S ∩ U 6= ∅, and
S ∩ W 6= ∅; thus, Uc ∩ Wc is an open subset of Sc (X)
− − e which is disjoint from
F and contains S.
From the previous paragraph we deduce that Y ⊆ F. To prove the reverse
containment let S ∈ F be arbitrary, set a = lim S, and fix W ∈ U ∈ C(X) e with
S ∈ hU ic and a ∈ W . Since a ∈ Vω , there is k < ω in such a way that
[
S \ {Vi : i ∈ k ∪ {ω}} ⊆ W. (10.1)

Now define U ∗ = {U ∈ U : U ∩ S ∩ Vω \ W 6= ∅} and for each U ∈ U ∗ choose

a point (ω, xU ) ∈ U ∩ S \ W . Use the fact that U ∗ is finite to get an integer
m ≥ k with
((ω + 1) \ m) × {xU } ⊆ U,
for all U ∈ U ∗ .
Finally, let e : U ∗ → ω \ m be an arbitrary one-to-one function and observe
that by letting
 [ 
g = S \ {Vi : i ∈ ran(e) ∪ {ω}} ∪ {(e(U ), xU ) : U ∈ U ∗ } ∪ {a}

we obtain that g ∈ Y ∩ hU ic .
To complete the proof of (1), let us show that F has void interior. If S ∈ F
and W ∈ U ∈ C(X) e satisfy that S ∈ hU ic and lim S ∈ W , then there is k < ω
such that (10.1) holds. In particular S \ Vk+1 ∈ hU ic \ F, in other words, the
interior of F is empty.
For the rest of the argument, let f ∈ Y be arbitrary. Notice that a base for
Sc (X)
e at f is given by all sets of the form

h{{i} × Ui : i < m} ∪ {((ω + 1) \ m) × U }ic , (?)

 CARDINAL FUNCTIONS OF Sc (X) 23

where m ∈ ω, f [(ω + 1) \ m] ⊆ U ∈ τX , and f (i) ∈ Ui ∈ τX , for each i < m.

Also, when G is of the form (?), we deduce that
\
G∩Y = [{i}, Ui ] ∩ [(ω + 1) \ m, U ]
i<m

and therefore, the relative topology of Y is coarser than the compact-open topol-
ogy.
Assume that U ∈ τX , K ∈ K (ω + 1), and f ∈ Y ∩ [K, U ] are arbitrary. If
ω∈/ K, then K is a finite subset of ω and
−
\
f ∈Y ∩ ({i} × U )c ⊆ [K, U ].
i∈K

On the other hand, ω ∈ K implies the existence of k < ω in such a way that
f [(ω + 1) \ k] ⊆ U ; hence,
V = {((ω + 1) \ k) × U } ∪ {{i} × U : i ∈ K ∩ k} ∪ {Vi : i ∈ k \ K}
e satisfying that f ∈ hV ic ∩ Y ⊆ [K, U ]. In
is a finite cellular family in X
conclusion, (2) is proved. 
Let us recall that the Lindelöf number of a topological space X (in symbols,
L(X)) is the least cardinal κ for which each open cover of X contains a subcover
of size ≤ κ. It is immediate that L(X) ≤ w(X).
Assume that X is a space with a nontrivial convergent sequence and consider
Figure 1, where the symbol κ → λ means λ ≤ κ and n ∈ N is arbitrary.
The rest of this section is devoted to the verification of all inequalities shown
in the figure.

w(X) - L(S (X))

c H
H HH
HH H
Hj
H HH
j
L(C(X))
e - L(X)
*

 
?
L(Sc (X))
e - L(X n ) 

Figure 1.

Let us start by verifying that w(X) is above all cardinals appearing in Fig-
ure 1. From the assumption, we get that w(X) = w(Sc (X)) ≥ L(Sc (X)) (see
Lemma 3.3). Another consequence of our hypothesis is that Sc (X) e 6= ∅ and so,
w(X) = w(X) = w(Sc (X)) ≥ L(Sc (X)). Finally, Theorem 10.3 and Lemma 3.3
e e e
give w(X) = w(X) e = w(Sc (X))e ≥ w(C(X))
e ≥ L(C(X)).
e
24 D. MAYA, P. PELLICER-COVARRUBIAS, AND R. PICHARDO-MENDOZA

Since any space containing a nontrivial convergent sequence contains a closed

copy of the linearly ordered space ω + 1, we deduce that if Sc (X) 6= ∅, then
L(X) ≥ ω and L(Sc (X)) ≥ ω (see Proposition 3.5).
Proposition 10.4. If X is a space with Sc (X) 6= ∅, then L(X) ≤ L(Sc (X)).
Proof. Set κ = L(S Sc (X)) and let U be an open cover of X.
Define U ∗ = { V : V ∈ [U ]<ω } and fix S ∈ Sc (X). If a = lim S, then
there is Ua ∈ U with a ∈ UaSand, also, for each x ∈ S \ Ua there exists Ux ∈ U
such that x ∈ Ux ; thus, U = {Ux : x ∈ (S \ Ua ) ∪ {a}} is a member of U ∗ with
S ∈ Uc+ . In other words, {Uc+ : U ∈ U ∗ } is an open cover of Sc (X).
Let U0 ∈ [U ∗ ]≤κ be so that {Uc+ : U ∈ U0 } covers Sc (X) S and, for each
U ∈ U0 , assume Sthat VU is a finite subset of U satisfying U = VU .
Since κ ≥ ω, {VU : U ∈ U0 } is a subset of U having at most κ elements.
To prove that it covers X, assume that S ∈ Sc (X) is arbitrary and note that
+
for each x ∈ X there exists U ∈ U0 with S ∪ {x} ∈ Uc+ = ( VU )c ; thus, for
S
some V ∈ VU , x ∈ V . 

Theorem 10.5. If n ∈ N and X is a nonempty space, then X n embeds as a

closed subspace of Sc (X).
e In particular, L(X n ) ≤ L(Sc (X)).
e

Proof. Set A = {{i} × X : i < n}. Fix a ∈ X and let S = ((ω + 1) \ n) × {a}.
Observe that A S ∈ [C L (X)]
e <ω is a pairwise disjoint family and S ∈ Sc (X)e
satisfies that S ∩ A = ∅. By Lemma 3.6, the set E1 (A , S) is a closed subset
of Sc (X)
e which is homeomorphic to Q
i<n F1 ({i} × X). It is known that each
F1 ({i} × X) is homeomorphic to X. So, E1 (A , S) is homeomorphic to X n . 

Let us note that Sc (X)

e 6= ∅ for any nonempty space X.

Theorem 10.6. If X is a space, then X embeds as a closed subspace of C(X).

e
In particular L(X) ≤ L(C(X)).
e

Proof. Define h : X → C(X)

e by h(x) = (ω + 1) × {x}, for each x ∈ X. Notice
that h is one-to-one. Now, given K ∈ K (ω + 1) and U ∈ τX , observe that
h−1 [[K, U ]] = U and h[U ] = [ω + 1, U ] ∩ h[X], i. e., h is an embedding.
It remains to show that h[X] is closed in C(X):
e if f ∈ C(X)
e \ h[X], then f
is not constant and so there are m < n ≤ ω in such a way that f (m) 6= f (n).
Fix two disjoint sets U, V ∈ τX with f (m) ∈ U and f (n) ∈ V , and note that

f ∈ [{m}, U ] ∩ [{n}, V ] ⊆ C(X)

e \ h[X].


Recall that S denotes Sorgenfrey’s line.

 CARDINAL FUNCTIONS OF Sc (X) 25

Example 10.7. Assume that κ is a non-zero cardinal with κ ≤ c and set

X = Sκ .
Let us argue that L(Sc (X)) = e(Sc (X)) = c. First, since Sc (X) 6= ∅, by
Lemma 3.3, we obtain that L(Sc (X)) ≤ w(Sc (X)) = w(X) = c. Now observe
that, by (2) of Example 6.3 and the fact that X contains a closed copy F of S,
we get that Fc+ is a closed subset of Sc (X) and c ≤ e(Sc (S)) ≤ e(Sc (X)) ≤
L(Sc (X)) (see [1, Problem 3.12.7(e), p. 225]), which completes our argument.
Also note that if n ∈ N and the product κn is equal to 1, then X n = S and
so, L(X n ) = L(S) = ω. On the other hand, when κn ≥ 2, X n contains a closed
copy of S2 and thus, c = L(S2 ) ≤ L(X n ) ≤ c, i.e., L(X n ) = c.
In connection with Proposition 10.4, note that Example 10.7 implies the fol-
lowing:
L(S) = ω < c = L(Sc (S)) and L(S2 ) = c = L(Sc (S2 )).
e κ )), when 2 ≤ κ ≤ c, but first we need a prelim-
Now we will calculate L(C(S
inary result.
Lemma 10.8. Suppose X and Y are topological spaces. If h : Y → X is an
e ) → C(X)
embedding, then the function h∗ : C(Y e given by h∗ (f ) = h ◦ f , for
each f ∈ C(Y
e ), is an embedding. Moreover, when ran(h) is a closed subset of
X, ran(h∗ ) is a closed subset of C(X).
e

Proof. It is straightforward to show that h∗ is one-to-one and that the equality

h∗ −1 [[K, V ]] = [K, h−1 [V ]] holds for all K ∈ K (ω + 1) and V ∈ τX . To finish
showing that h∗ is an embedding, let U ∈ τY and K ∈ K (ω +1) be arbitrary; fix
V ∈ τX with h[U ] = V ∩ ran(h) and observe that h∗ [[K, U ]] = [K, V ] ∩ ran(h∗ ).
For the second statement in our lemma: assume f ∈ C(X)\ran(h
e ∗ ). If ran(f )
−1
were a subset of ran(h), then h ◦ f would be a member of C(Y ) which is sent
e
by h∗ to f , contradicting our hypothesis. Fix i ≤ ω such that f (i) ∈
/ ran(h) and
note that [{f (i)}, X \ ran(h)] is an open subset of C(X)
e which contains f and
is disjoint from ran(h∗ ). 
As a consequence of this result: when 2 ≤ κ ≤ c, S2 embeds as a closed
subspace of X = Sκ and hence, C(X)
e e 2 ). On the
contains a closed copy of C(S
other hand, Theorem 10.6 guarantees that there is a closed copy of S2 in C(X)
e
2
and therefore, c = L(S ) ≤ L(C(X)) ≤ c, i.e., L(C(X)) = c.
e e

Example 10.9. There are spaces X for which

w(X) = L(X n ) = L(C(X))
e = L(Sc (X))
e = L(Sc (X)) = c > ω = L(X),
whenever n ∈ N.
Question 10.10. Separate (if possible) all cardinals from Figure 1, i.e., for each
pair, f and g, of cardinal functions appearing in that figure and satisfying
26 D. MAYA, P. PELLICER-COVARRUBIAS, AND R. PICHARDO-MENDOZA

f (X) ≤ g(X), find a space Y such that f (Y ) 6= g(Y ). For example, if f (X) =
L(Sc (X)) and g(X) = L(X), then, in Example 10.9, we found a space Y with
f (Y ) > g(Y ).
Question 10.11. Assume that f ∈ {c, t, e, ∆, nw, iw, πχ} is a cardinal function.
Find necessary and sufficient conditions on a space X to get the equality f (X) =
f (Sc (X)).

References
[1] ENGELKING, R., General Topology, Sigma Series in Pure Mathematics, vol. 6, Helder-
mann Verlag, Berlin, 1989, Translated from Polish by the author.
[2] FEDORCHUK, V. — TODORČEVIĆ, S., Cellularity of covariant functors, Topology
Appl. 76 (1997), 125–150.
[3] GARCÍA-FERREIRA, S. — ORTIZ-CASTILLO, Y. F., The hyperspace of convergent
sequences, Topology Appl. 196 (2015), 795–804.
[4] GRUENHAGE, G. — TANAKA, Y., Products of k-spaces and spaces of countable tight-
ness, Trans. Amer. Math. Soc. 273 (1982), no. 1, 299–308.
[5] KUNEN, K., Set Theory. An introduction to independence proofs, vol. 102, North-Holland
Publishing Co., Amsterdam, 1980.
[6] MAYA, D. — PELLICER-COVARRUBIAS, P. — PICHARDO-MENDOZA, R., General
properties of the hyperspace of convergent sequences, submitted.
[7] MICHAEL, E., Topologies on spaces of subsets, Trans. Amer. Math. Soc. 71 (1951),
152–182.

* Departamento de Matemáticas,
Facultad de Ciencias, UNAM,
Circuito ext. s/n, Ciudad Universitaria,
C.P. 04510,
México, D.F.
E-mail address: dmayae@outlook.com

** Departamento de Matemáticas,
Facultad de Ciencias, UNAM,
Circuito ext. s/n, Ciudad Universitaria,
C.P. 04510,
México, D.F.
E-mail address: paty@ciencias.unam.mx

*** Departamento de Matemáticas,

Facultad de Ciencias, UNAM,
Circuito ext. s/n, Ciudad Universitaria,
C.P. 04510,
México, D.F.
E-mail address: rpm@ciencias.unam.mx
URL: http://www.matematicas.unam.mx/pmr