7

1
4


r
e
v
i
s
i
o
n
:
2
0
0
1
-
0
4
-
0
2







m
o
d
i
f
i
e
d
:
2
0
0
1
-
0
4
-
0
2


A TALL SPACE WITH A SMALL BOTTOM
ISTV

AN JUH

ASZ, SAHARON SHELAH, LAJOS SOUKUP, AND ZOLT

AN
SZENTMIKL

OSSY
Abstract. We introduce a general method of constructing locally compact
scattered spaces from certain families of sets and then, with the help of this
method, we prove that if
<
= then there is such a space of height
+
with only many isolated points. This implies that there is a locally compact
scattered space of height
2
with
1
isolated points in ZFC, solving an old
problem of the rst author.
1. Introduction
Let us start by recalling that a topological space X is called scattered if ev-
ery non-empty subspace of X has an isolated point and that such a space has a
natural decomposition into levels, the so called Cantor-Bendixson levels. The
th
Cantor-Bendixson level of X will be denoted by I

(X). We shall write I

<
(X) =

<
I

(X). The height of X, ht(X), is the least with I

(X) = . The sequence

[I

(X)[ : ht(X)) is said to be the cardinal sequence of X. The width of X,

wd(X), is dened by wd(X) = sup [ I

(X)[ : < ht(X).

The cardinality of a T
3
, in particular of a locally compact, scattered (in short:
LCS) space X is at most 2
| I(X)|
, hence clearly ht(X) < (2
| I(X)|
)
+
. Therefore under
CH there is no LCS space of height
2
with only countably many isolated points.
On the other hand, I. Juh asz and W. Weiss, [3, theorem 4], proved in ZFC that for
every <
2
there is a LCS space X with ht(X) = and wd(X) = . The natural
question if the existence of an LCS space of height
2
with countable width follows
from CH was answered in the negative by W. Just, who proved, [4, theorem 2.13
], that if one adds Cohen reals to a model of CH then in the generic extension there
are no LCS spaces of height
2
and wd(X) = . On the other hand, Baumgartner
and Shelah proved it consistent (with CH) that such an LCS space exists.
The above mentioned estimate ht(X) < (2
|I(X)
[)
+
is sharp for LCS spaces with
countably many isolated points : it is easy to construct an LCS space with countable
bottom and of height for each < (2

)
+
(see theorem 2.20). Much less is
known about LCS spaces with
1
isolated points, for example it is a long standing
open problem whether there is, in ZFC, an LCS space of height
2
and width
1
. In
fact, as was noticed by Juh asz in the mid eighties, even the much simpler question
1991 Mathematics Subject Classication. 54A25, 06E05, 54G12, 03E20.
Key words and phrases. locally compact scattered space, superatomic Boolean algebra.
The rst, third and fourth authors were supported by the Hungarian National Foundation for
Scientic Research grant no. 25745 .
The second author was supported by the United States Israel Binational Science Foundation,
Publication 714.
The third author was partially supported by Grant-in-Aid for JSPS Fellows No. 98259 of the
Ministry of Education, Science, Sports and Culture, Japan.
1
7
1
4


r
e
v
i
s
i
o
n
:
2
0
0
1
-
0
4
-
0
2







m
o
d
i
f
i
e
d
:
2
0
0
1
-
0
4
-
0
2


2 I. JUH

ASZ, S. SHELAH, L. SOUKUP, AND Z. SZENTMIKL

OSSY
if there is a ZFC example of an LCS space of height
2
with only
1
isolated points,
turned out to be surprisingly dicult. On the other hand, Martnez, [6, theorem 1]
proved that it is consistent that for each <
3
there is a LCS space of height
and width
1
. As the main result of the present paper, we shall give an armative
answer to the above question of Juh asz: in section 2 we construct, in ZFC, an LCS
space of height
2
with
1
isolated points. Since this space we construct in theorem
2.21 has width
2
, the following question remains:
Problem 1. Is there an LCS space X of height
2
and width
1
in ZFC?
The methods used in the proof of theorem 2.21 do not seem to suce to get LCS
spaces with
1
isolated points of arbitrary height <
3
. Thus we have the following
problem:
Problem 2. Is there, in ZFC, an LCS space with
1
isolated points and of height
for each <
3
?
Although one of our main results, theorem 2.19, generalizes to higher cardinals,
it does not seem to suce to get the analogous result e.g. for
3
instead of
2
. If
2

2
but 2
1
>
2
then neither theorem 2.20 nor theorem 2.19 can be applied
to get an LCS space of height
3
with only
2
many isolated points. Thus the
following version of Juh asz problem remains open:
Problem 3. Is there, in ZFC, an LCS space X of height
3
having
2
isolated
points?
Let us mention here that the problem of the existence of (
+
, )-thin-tall spaces,
i. e. LCS spaces of width and height
+
, is mentioned in [9, Problem 6.4, p.53].
However, it is erroneously stated there that the existence of a (
+
, )-thin-tall space
follows from
<
= or from the existence of a
+
-tree.
2. A space of height
2
and with
1
isolated points
Denition 2.1. Given a family of sets / we dene the topological space X(/) =
/,
A
) as follows:
A
is the coarsest topology in which the sets U
A
(A) = /T(A)
are clopen for each A /, in other words: U
A
(A), / U
A
(A) : A / is a
subbase for
A
.
We shall write U(A) instead of U
A
(A) if / is clear from the context.
Clearly X(/) is a 0-dimensional T
2
-space. A family / is called well-founded i
/, ) is well-founded. In this case we can dene the rank-function rk : / On
as usual:
rk(A) = suprk(B) + 1 : B A,
and write R

(/) = A / : rk(A) = .
The family / is said to be -closed i A B / whenever A, B /.
It is easy to see that if / is -closed, then a neighbourhood base in X(/) of
A / is formed by the sets
W(A; B
1
, . . . , B
n
) = U(A)
n
_
1
U(B
i
),
where n and B
i
A for i = 1, . . . , n. ( For n = 0 we have W(A) = U(A).)
The following simple result enables us to obtain LCS spaces from certain families
of sets. Let us point out, however, that not every LCS space is obtainable in this
manner, but we do not dwell upon this because we will not need it.
7
1
4


r
e
v
i
s
i
o
n
:
2
0
0
1
-
0
4
-
0
2







m
o
d
i
f
i
e
d
:
2
0
0
1
-
0
4
-
0
2


A TALL SPACE WITH A SMALL BOTTOM 3
Lemma 2.2. Assume that / is both -closed and well-founded. Then X(/) is an
LCS space.
Proof. Given a non-empty subset } of / let A be a -minimal element of }. Then
U(A) } = A, i.e. A is isolated in }. Thus X(/) is scattered.
Next we prove that every U(A) is compact by well-founded induction on /, ).
Assume that U(B) is compact for each B A. By Alexanders subbase lemma it
is enough to prove that any cover of U(A) with subbase elements contains a nite
subcover. So let B, ( / be such that
U(A)
_
BB
U(B)
_
CC
(/ U(C)).
If A U(B) for some B B then A B and so U(A) U(B).
Hence we can assume that A / U(C), i.e. A , C for some C (. If
A C = then U(A) / U(C), and we are clearly done. So we can
assume that A C ,= , and consequently A C /. Then U(A) (/ U(C)) =
U(A) U(C) = U(A C). Since A C ,= A the set U(A C) is compact by
the induction hypothesis, hence U(A C) is covered by a nite subfamily T of
U(B) : B B / U(D) : D (. Therefore T / U(C) is a nite cover
of U(A). Consequently U(A) is compact.
To simplify notation, if X(/) is scattered then we write I

(/) = I

(X(/)).
Clearly each minimal element of A / is isolated in X(/); more generally we
have rk(A) if A I

(/), as is shown by an easy induction on rk(A).

Example 2.3. Assume that T, ) is a well-ordering, tp T, ) = , and let / be
the family of all initial segments of T, ), i. e. / = T T
x
: x T, where
T
x
= t T : t x. Then / is well-founded, -closed and it is easy to see that
X(/)

= +1, i.e. the space X(/) is homeomorphic to the space of ordinals up to
and including .
Example 2.3 above shows that, in general, R

(/) and I

(/) may dier even for

= 0. Indeed, if x is the successor of y in T, ) then T
x
is isolated in X(/)
because T
x
= W(T
x
; T
y
) = U
A
(T
x
) U
A
(T
y
) is open, but rk(x) = tp(T
x
) > 0.
However, for a wide class of families, the two kinds of levels do agree. Let us call a
well-founded family / rk-good i the following condition is satised:
A / < rk(A) [A

/ : A

A rk(A

) = [ .
Then we have the following result.
Lemma 2.4. If / is well-founded, -closed and rk-good then I

(/) = R

(/) for
each .
Proof. We prove this by induction on . Assume that I

(/) = R

(/) for all < .

If A R

(/) then U(A) A

<
R

(/) =

<
I

(/) and so A is an
isolated point of /

<
I

(/), i.e. A I

(/). Thus we have R

(/) I

(/).
Now assume that A I

(/) R

(/). Then by our above remark < rk(A),

moreover there are B
1
, . . . B
n
U
A
(A) A such that
() W(A; B
1
, . . . , B
n
) A I
<
(/) = R
<
(/).
Let = max, max
i=1,...,n
rk B
i
. Then < rk(A), moreover we have U
A
(A)
R

(/) B
1
, . . . B
n
by (), contradicting [ U
A
(A) R

(/)[ . Note that this

7
1
4


r
e
v
i
s
i
o
n
:
2
0
0
1
-
0
4
-
0
2







m
o
d
i
f
i
e
d
:
2
0
0
1
-
0
4
-
0
2


4 I. JUH

ASZ, S. SHELAH, L. SOUKUP, AND Z. SZENTMIKL

OSSY
argument is valid for n = 0 as well. Indeed, in this case we have U
A
(A) A
I
<
(/), moreover = . Thus we have concluded that I

(/) = R

(/).
Example 2.5. For a xed cardinal and any ordinal <
+
we dene the family
c

T(

) as follows:
c

=
_
_

1+
,
1+
( + 1)
_
: ,
1+
<

_
.
Of course, throughout this denition exponentiation means ordinal exponentia-
tion.
c

is clearly well-founded, -closed, moreover rk

_
_

1+
,
1+
( +1)
_
_
= ,
hence c

is also rk-good. Consequently X(c

) is an LCS space of height + 1 in

which the
th
level is
_
_

1+
,
1+
( + 1)
_
:
1+
<

_
, i. e. all levels
except the top one are of size .
To get an LCS space of height
+
with few isolated points, our plan is to
amalgamate the spaces X(c

) : <
+
into one LCS space X in such a way
that [ I
0
(X)[
<
. The following denition describes a situation in which such
an amalgamation can be done.
Denition 2.6. A system of families /
i
: i I is called coherent i A B
/
i
whenever i, j
_
I

2
, A /
i
and B /
j
.
To simplify notation, we introduce the following convention. Whenever the sys-
tem of families /
i
: i I is given, we will write U
i
(A) for U
Ai
(A), and
i
for
Ai
.
If the family / is dened then we will write U(A) for U
A
(A), and for
A
.
Lemma 2.7. Assume that /
i
: i I is a coherent system of well-founded, -
closed families and / = /
i
: i I. Then for each i I and A /
i
we have
U(A) = U
i
(A), / is also well-founded and -closed, moreover
i
[`U(A) = [`U(A).
Consequently each X(/
i
) is an open subspace of X(/) and thus X(/
i
) : i I
forms an open cover of X(/).
Proof. Let A /
i
. Then it is clear from coherence that
U
i
(A) U(A) =
_
jI
B : B /
j
B A U
i
(A),
hence U
i
(A) = U(A).
Next let B /
j
. If A B = then U(A) U(B) . Now assume that
A B ,= . Then, again by coherence, A B /
i
and we have
U(A) U(B) = U
i
(A) U(B) = C /
i
: C A C B
= C /
i
: C A B = U
i
(A B).
In both cases U(A) U(B) is
i
-open. Similarly we can see that U(A) U(B) =
U
i
(A) U(B) is
i
-open, hence the topologies
i
[`U(A) and [`U(A) coincide.
To show that / is well-founded, assume that A
n
: n / and A
0
A
1

. . . . If A
0
/
i
then A
n
: n /
i
because U(A
0
) = U
i
(A
0
). Thus there is
n with A
m
= A
n
for each m n because /
i
is well-founded. Finally, that /
is -closed is an easy consequence of coherence and the -closednessof the families
/
i
.
7
1
4


r
e
v
i
s
i
o
n
:
2
0
0
1
-
0
4
-
0
2







m
o
d
i
f
i
e
d
:
2
0
0
1
-
0
4
-
0
2


A TALL SPACE WITH A SMALL BOTTOM 5
Given a system of families /
i
: i I we would like to construct a coherent
system of families

/
i
: i I such that /
i
and

/
i
are isomorphic for all i I. A
sucient condition for when this can be done will be given in lemma 2.9 below.
First, however, we need a denition. While reading it, one should remember
that an ordinal is identied with the family of its proper initial segments.
Denition 2.8. Given a limit ordinal and a family / with / T(), let us
dene the family

/ as follows. Consider rst the function k
A
on determined by
the formula k
A
() = U
A
( + 1) for and put

/ = k

A
A : A /.
Since /, for each we clearly have U
A
() = and so k
A
() =
U
A
( + 1) ,= U
A
( + 1) = k
A
() whenever ,
_

2
. Consequently, k
A
is a
bijection that yields an isomorphism between / and

/ (and so the spaces X(/)
and X(

/) are homeomorphic).
If the system of families /
i
: i I is given, then we write k
i
for k
Ai
for each
i I.
If / T() and then we let
/[` = A : A /.
For /
0
,= /
1
T() we let
(/
0
, /
1
) = min : /
0
[` ,= /
1
[`.
Clearly we always have (/
0
, /
1
) . If, in addition, +1 /
0
/
1
, moreover
both /
0
and /
1
are -closed then we also have
(/
0
, /
1
) = min : U
0
() ,= U
1
(),
because then /
i
[` = U
i
() whenever i 2 and .
Lemma 2.9. Assume that is a cardinal, /
i
: i I TT() are -closed
families, + 1 /
i
for each i I, and (/
i
, /
j
) is a successor ordinal whenever
i, j
_
I

2
. Then the system

/
i
: i I is coherent.
Proof. Let A /
i
and B /
j
, where (/
i
, /
j
) = + 1. Then B U
j
() =
U
i
() by the choice of and so A B /
i
because /
i
is -closed.
Consequently we have
k

i
A k

j
B =
_
U
i
( + 1) : A B U
i
( + 1) = U
j
( + 1)
_
=
_
U
i
( + 1) : A B <
_
= k

i
(A B )

/
i
,
as required by the denition of coherence.
More is needed still if we want the amalgamated family to provide us a space
with a small base, i.e. having not too many isolated points. This will be made clear
by the following lemma.
Lemma 2.10. Let be a cardinal and /
i
: i I TT() be a system of families
such that
(i)
.
+ 1 /
i
and /
i
is well-founded and -closed for each i I,
(ii) (/
i
, /
j
) is a successor ordinal for each i, j
_
I

2
.
7
1
4


r
e
v
i
s
i
o
n
:
2
0
0
1
-
0
4
-
0
2







m
o
d
i
f
i
e
d
:
2
0
0
1
-
0
4
-
0
2


6 I. JUH

ASZ, S. SHELAH, L. SOUKUP, AND Z. SZENTMIKL

OSSY
Then
(a) the system

/
i
: i I is coherent and thus / =

/
i
: i I is well-founded,
-closed and X(/) is covered by its open subspaces X(

/
i
) : i I.
If, in addition, we also have
(iii) I
0
(/
i
)
_

<
for each i I,
and
(iv) [ U
i
()[ < for each i I and ,
then
(b) I
0
(/)[
_
_
[]
<
_
<
_
<
.
Proof of lemma 2.10. The system

/
i
: i I is coherent by lemma 2.9, thus (a)
holds by lemma 2.7.
Consequently we have
I
0
(/) =
_
I
0
(

/
i
) : i I.
Now if A I
0
(/
i
) and then [A[ < by (iii) and U
i
()
_
_

<
_
<
by
(iv), hence
k

i
A = U
i
( + 1) : A
_
_
_

<
_
<
_
<
.
This, by I
0
(

/
i
) = k

i
A : A I
0
(/
i
), proves (b).
Before we could apply this result to the families c

, however, we need some

further preparation.
Denition 2.11. A family / is called tree-like i A A

,= implies that A A

or A

A, whenever A, A

/.
Denition 2.12. A family / is called chain-closed if for each non-empty B /
if B is ordered by (i.e. if B is a chain) then B /.
It is easy to see that the families c

given in example 2.5 are both tree-like and

chain-closed. Also, tree-like families are clearly -closed.
Lemma 2.13. If is an ordinal and / T() is tree-like, well-founded and chain-
closed then so is /[` for each .
Proof of lemma 2.13. It is obvious that /[` is tree-like. To show that /[` is
chain-closed, let ,= B /[` be ordered by . If B = then B = /[`.
If, however, B ,= then put

B = A / : A B . Since / is tree-like,

B is also ordered by and clearly

B ,= . So

B / and B =

B /[`,
which was to be shown.
To show that /[` is well-founded assume that A
0
A
1
. . . , where
each A
n
/. If A
n
= for some n, then we are done. Otherwise for each
n we have A
n
= (
mn
A
m
) ,= , hence as / is -closed we can assume
that A
0
A
1
. . . . Since / is well-founded, there is n such that A
m
= A
n
, and
so A
m
= A
n
as well, for each m n.
7
1
4


r
e
v
i
s
i
o
n
:
2
0
0
1
-
0
4
-
0
2







m
o
d
i
f
i
e
d
:
2
0
0
1
-
0
4
-
0
2


A TALL SPACE WITH A SMALL BOTTOM 7
Denition 2.14. Given a family / T() and , let us put
S
A
(, ) = A / : A and / A.
Lemma 2.15. Assume that is an innite ordinal and / T() is a tree-like,
well-founded and chain-closed family with /. Then
/ = S
A
(, ) : , .
Consequently, [/[ [[.
Proof of lemma 2.15. Given , , the family o = A / : A, / A is
ordered by because / is tree-like. Thus either o = and so S
A
(, ) = o = ,
or if o ,= then S
A
(, ) = o /, for / is chain-closed.
Assume now that A / , and let T = D / : A D. Clearly
T. Since / is tree-like, T is ordered by , so it has a -least element, say
D, because /, ) is also well-founded. Pick D A and let A. We claim
that A = S
A
(, ). Clearly A S
A
(, ) because A and / A. On the other
hand, if A

/, A

and / A

then either A

A or A A

because / is
tree-like. But / A

implies that A

/ T, i.e. A A

can not hold. Thus A

A
and so S
A
(, ) = A is proved.
Denition 2.16. If is an ordinal and / T() let us put
/

= A : A / = / A : A / < .
Lemma 2.17. If is an ordinal and /
0
, /
1
T() are chain-closed, -closed and
well-founded families such that /
0

,= /
1

then (/
0

, /
1

) is a successor ordinal.
Proof. Assume that is a limit ordinal and /
0

[` = /
1

[` for all < . We

want to show that /
0

[` = /
1

[`. Since /
i

[` =

/
i
[` and

<
/
i
[` =

<
/
i

[`, moreover /
0

/
1

, it is enough to show that (/

0
[`) =
(/
1
[`) .
So assume that A /
0
with A ,= and verify that then A /
1
[`.
Fix A . For each with < < let B

be the -minimal element of

/
1
with B

= A ,= . Then B

: < is a chain because B

for
<

by the minimality of B

and because /
1
is -closed. Thus B = B

: <
< /
1
and clearly A = B .
The last result shows us that the operation * is useful because its application
yields us families that satisfy condition (ii) of lemma 2.10. On the other hand, the
following result tells us that the LCS spaces associated with the families modied by
* do not dier signicantly from the spaces given by the original families, moreover
they also satisfy condition (iii) of lemma 2.10.
Lemma 2.18. Let be a cardinal and /
_

be well-founded and -closed.

Then so is /

, moreover
(a) X(/) is a closed subspace of X(/

),
(b) I
0
(/) I

(/

).
(c) ht(/

)
.
+ ht(/),
(d) I
0
(/

)
_

<
.
7
1
4


r
e
v
i
s
i
o
n
:
2
0
0
1
-
0
4
-
0
2







m
o
d
i
f
i
e
d
:
2
0
0
1
-
0
4
-
0
2


8 I. JUH

ASZ, S. SHELAH, L. SOUKUP, AND Z. SZENTMIKL

OSSY
Proof of lemma 2.18. We shall write U(A) for U
A
(A), and U

(A) for U
A
(A).
First observe that because
U

(A) / =
_
U(A) if A /,
if A /

/,
X(/) is a closed subspace of X(/

), hence (a) holds.

Now let A I
0
(/). Then there are B
1
, . . . , B
n
U(A) A such that
A = W(A; B
1
, . . . , B
n
) = U(A)
n
_
i=1
U(B
i
).
Since here B
i
A and [A[ = , we can x A such that (A ) , B
i
for every
i = 1, . . . , n.
Now consider the basic neighbourhood
Z = W

(A; A , B
1
, . . . , B
n
) = U

(A) U

(A )
n
_
i=1
U

(B
i
)
of A in X(/

). We claim that Z = A : < . The inclusion is clear

from the choice of . On the other hand, if C Z with C / and , then
C A hence C = AC , so as / is -closed we can assume that C A.
If we had C ,= A then A = W(A; B
1
, . . . , B
n
) would imply C B
i
for some i,
hence C U

(B
i
) and so C / Z, a contradiction, thus we must have C = A.
Moreover, since U

(A ) A : , we must also have > .

By example 2.3 we have X(Z)

= +1. Moreover, the topologies
Z
and
A
[`Z
coincide because the above argument also shows that for each C / and we
have
U

(C ) Z = U

(A C ) Z =
_
U
Z
(A ) if A C and > ;
otherwise.
Hence X(Z)

=
.
+ 1 is a clopen subspace of X(/

) and so A = I

(Z) =
I

(/

) Z, what proves (b).

(c) follows immediately from (a) and (b).
Finally, I
0
(/

) I
<
(/

) (/

/)
_

<
, as follows immediately from (b),
proving (d).
Now we are ready to collect the fruits of all the preparatory work.
Theorem 2.19. If
<
= then there is an LCS space X of height
+
with
[ I
0
(X)[ = .
Proof of theorem 2.19. For each <
+
consider the well-founded, -closed, rk-
good family c

constructed in example 2.5:

c

=
_
_

1+
,
1+
( + 1)
_
: ,
1+
<

_
.
Fix a bijection f

, and let T

= f

E : E c

, i.e. T

is simply
an isomorphic copy of c

on the underlying set . As c

is also chain-closed and

tree-like, hence so is T

.
We shall now show that the *-modied families T

: <
+
satisfy conditions
(i)-(iv) of lemma 2.10. Since T

it follows that
.
+ 1 T

and so (i) is true.

For ,
_

2
, the height of X(c

) is + 1 and the height of X(c

) is + 1,
hence c

and c

are not isomorphic. Thus T

,= T

and so T

,= T

because
7
1
4


r
e
v
i
s
i
o
n
:
2
0
0
1
-
0
4
-
0
2







m
o
d
i
f
i
e
d
:
2
0
0
1
-
0
4
-
0
2


A TALL SPACE WITH A SMALL BOTTOM 9
T

= T

and T

= T

. Hence (T

, T

) is a successor ordinal
by lemma 2.17, i.e. (ii) is satised.
(iii) holds by 2.18.(d.)
To show (iv), let us x < . Then U
F
() = T

[` = T

[` : where
[T

[`[ [[
2
for all by lemmas 2.13 and 2.15, consequently [T

[`[ [[
3
<
.
Thus we may apply lemma 2.10 to the family T =

: <
+
and
conclude that the space X = X(T) is LCS, [ I
0
(X)[
_
(
<
)
<
_
<
= , moreover
since for every
+
the space X(T

) is an open subspace of X, we have

ht(X) ht(X(T

)) > , consequently ht(X)

+
.
In particular, if 2

=
1
then the above result yields an LCS space X with
ht(X) =
2
and [ I
0
(X)[ =
1
. That such a space also exist under CH, hence in
ZFC, follows from the following result.
Theorem 2.20. For each < (2

)
+
there is a locally compact, scattered space X

with [X

[ [[ + , ht(X

) = and [ I
0
(X

)[ = .
Proof. We do induction on . If = +1 then we let X

be the 1-point compact-

ication of the disjoint topological sum of countably many copies of X

.
If is limit then we rst x an almost disjoint family A

: <
_

, for
[[ 2

. Applying the inductive hypothesis for each < we also x a locally

compact scattered space X

of height such that I

0
(X

) = A

and X

=
A

for ,
_

2
. Now amalgamate the spaces X

as follows: consider the

topological space X =
<
X

, ) where is the topology generated by

<

.
Since A

is a nite and open subspace of both X

and X

it follows that each

X

is an open subspace of X. Consequently, X is LCS with countably many isolated

points, and ht(X) = sup
<
ht X

= .
Corollary 2.21. There is a locally compact, scattered space of height
2
and having

1
isolated points.
Proof. If 2

=
1
, then theorem 2.19 gives such a space.
If 2

>
1
then (2

)
+

3
and so according to theorem 2.20 for each <
3
there is locally compact, scattered space of height and countably many isolated
points.
References
[1] J. E. Baumgartner, S. Shelah, Remarks on superatomic Boolean algebras, Ann. Pure Appl.
Logic, 33 (l987), no. 2, 109-129.
[2] A. Dow, Large compact separable spaces may all contain N, Proc. Amer. Math. Soc.
109(1990) no.n1. 275279.
[3] I. Juhasz, W. Weiss, On thin-tall scattered spaces, Colloquium Mathematicum, vol XL (1978)
6368.
[4] W. Just, Two consistency results concerning thin-tall Boolean algebras Algebra Universalis
20(1985) no.2, 135142.
[5] K. Kunen, Set Theory, North-Holland, New York, 1980.
[6] J. C. Martnez, A forcing construction of thin-tall Boolean algebras, Fundamenta Mathemat-
icae, 159 (1999), no 2, 99-113.
[7] S. Mrowka, M. Rajagopalan, T. Soudararajan, A characterization of compact scattered spaces
through chain limits (chain compact spaces), Proceedings of TOPO 72, Pittsburg 1972, Lec-
ture Notes 378, 288297, Berlin 1974.
7
1
4


r
e
v
i
s
i
o
n
:
2
0
0
1
-
0
4
-
0
2







m
o
d
i
f
i
e
d
:
2
0
0
1
-
0
4
-
0
2


10 I. JUH

ASZ, S. SHELAH, L. SOUKUP, AND Z. SZENTMIKL

OSSY
[8] Judy Roitman, Height and width of superatomic Boolean algebras, Proc. Amer. Math. Soc.
94(1985), no 1, 914.
[9] S. Shelah On what I do not understand (and have something to say) Part I Fundamenta
Mathematicae, 166 (2000), no 1-2, pp 1-82.
Alfred Renyi Institute of Mathematics
E-mail address: [email protected]
The Hebrew University of Jerusalem
E-mail address: [email protected]
Alfred Renyi Institute of Mathematics
E-mail address: [email protected]
E otv os University of Budapest
E-mail address: [email protected]