i

i
i
i
i
i
i
i
Publ. RIMS, Kyoto Univ.
43 (2007), 425442
Structure of Quasi-analytic Ultradistributions
Dedicated to Professor Akira Kaneko on his sixtieth birthday
By
Takashi Takiguchi

Abstract
We study the structure of functions between distributions and hyperfunctions.
The structure theorem is known for distributions, non-quasi-analytic ultradistribu-
tions and hyperfunctions. In this paper, we try to ll the gap among them. We prove
the structure theorem for quasi-analytic ultradistributions.
1. Introduction
In this paper, we discuss the structure of generalized functions. It is well-
known that any distribution f is locally represented as f = P(D)g, where P(D)
is a nite order dierential operator with constant coecients and g is a contin-
uous function, which is the structure theorem for distributions. The structure
theorems for non-quasi-analytic ultradistributions ([1, 5]) and hyperfunctions
([3]) are also known. In this paper, we study the structure of functions between
them, namely, the structure of quasi-analytic ultradistributions. We prove the
structure theorem for non-analytic ultradistributions which includes both non-
quasi-analytic and quasi-analytic ones. It is our main theorem to prove that
any non-analytic ultradistribution f of the class is locally represented as
f = P(D)g, where P(D) is an ultradierential operator of the class and g
is an ultradierentiable function of the class > . We also claim that this
Communicated by T. Kawai. Received April 17, 2006.
2000 Mathematics Subject Classication(s): Primary 46F05; Secondary 46F12, 46F15.
Key words: structure of generalized functions

Department of Mathematics, National Defense Academy of Japan, 1-10-20 Hashirimizu,

Yokosuka, Kanagawa 239-8686, Japan.
e-mail: takashi@nda.ac.jp
c 2007 Research Institute for Mathematical Sciences, Kyoto University. All rights reserved. i
i
i
i
i
i
i
i
426 Takashi Takiguchi
ultradierentiable function g can be taken from any class satisfying > .
Our main theorem gives the structure theorem for quasi-analytic ultradistribu-
tions and the proof of our main theorem gives another proof of the structure
theorem for non-quasi-analytic ultradistributions. In the proof of our main
theorem, it is essentially important to construct ultradierential operators of
the given non-analytic class. In [5], H. Komatsu applied an innite product
P() :=

p=1
_
1 +
l
2
p

2
m
2
p
_
, where m
p
:=
M
p
M
p1
,
2
:=
2
1
+ +
2
n
and l
p
is
some sequence of positive numbers, to construct the symbol of an ultradieren-
tial operator in the given non-quasi-analytic class, which does not converge in
quasi-analytic classes. On the other hand, it is not easy to modify A. Kanekos
method in [3] to construct an ultradierential operator suitable for our purpose,
since the class of non-analyticity is strictly given in our theory. Therefore we
apply our original method to construct the symbols of ultradierential opera-
tors. Before proving our main theorem, we prepare some elementary properties
of quasi-analytic ultradistributions.
2. Ultradistributions
In this section, we review the denition of ultradistributions. Let R
n
be an open subset and M
p
, p = 0, 1, . . ., be a sequence of positive numbers.
For non-quasi-analytic classes, we impose the following conditions on M
p
.
(M.0) (normalization)
M
0
= M
1
= 1.
(M.1) (logarithmic convexity)
M
2
p
M
p1
M
p+1
, p = 1, 2, . . . .
(M.2) (stability under ultradierential operators)
G, H such that M
p
GH
p
min
0qp
M
q
M
pq
, p = 0, 1, . . . .
(M.3) (strong non-quasi-analyticity)
G such that

q=p+1
M
q1
M
q
Gp
M
p
M
p+1
, p = 1, 2, . . . .
(M.2) and (M.3) are often replaced by the following weaker conditions
respectively;
(M.2)

(stability under dierential operators)

G, H such that M
p+1
GH
p
M
p
, p = 0, 1, . . . . i
i
i
i
i
i
i
i
Structure of Ultradistributions 427
(M.3)

(non-quasi-analyticity)

p=1
M
p1
M
p
< .
For two sequences M
p
and N
p
of positive numbers we dene their orders.
Denition 2.1. Let M
p
and N
p
be the sequences of positive numbers.
(i) M
p
N
p
if and only if there exist such constants L > 0 and C > 0 that
M
p
CL
p
N
p
for any p.
(ii) M
p
N
p
if and only if for any L > 0 there exists such a constant C > 0
that M
p
CL
p
N
p
for any p.
In order to dene quasi-analytic classes, we impose the following condi-
tions, (QA) and (NA), instead of (M.3) or (M.3)

.
(QA) (quasi-analyticity)
p! M
p
,

p=1
M
p1
M
p
= .
Let M
p
be a sequence of positive numbers satisfying (QA). If
liminf
p
p

p!
M
p
> 0
then c
{M
p
}
is the class of analytic functions. We impose the condition that
M
p
would not dene the analytic class, namely,
(NA) (non-analyticity)
lim
p
p

p!
M
p
= 0.
Denition 2.2. Let M
p
be a sequence of positive numbers and R
n
be an open subset. A function f c() = C

() is called an ultradierentiable
function of the class (M
p
) (resp. M
p
) if and only if for any compact subset
K and for any h > 0 there exists such a constant C (resp. for any compact
subset K there exist such constants h and C) that
(1) sup
xK
[D

(x)[ Ch
||
M
||
for all
holds. Denote the set of the ultradierentiable functions of the class (M
p
) (resp.
M
p
) on by c
(M
p
)
() (resp. c
{M
p
}
()) and denote by T

() the set of all

functions in c

() with their supports compact in , where = (M

p
) or M
p
. i
i
i
i
i
i
i
i
428 Takashi Takiguchi
Let K R
n
be a compact set, M
p
satisfy (M.1) and (NA). Denote by
c

[K] the set of the ultradierentiable functions of the class = (M

p
) or M
p

dened on some neighborhood of K. We dene c

{M
p
},h
[K] if and only if
c
{M
p
}
[K] and (1) holds for given h > 0.
For M
p
satisfying (M.3)

and a compact subset K let

(2) T

K
= T

(R
n
) ; suppf K,
where = (M
p
) or M
p
and we dene
(3) T
{M
p
},h
K
= T
{M
p
}
K
; C such that sup
xK
[D

(x)[ Ch
||
M
||
.
Let M
p
satisfy (M.1) and (M.3)

. We dene T

() as the strong dual of T

()
for any open set and call it the set of ultradistributions of the class dened
on . These spaces are endowed with natural structure of locally convex spaces.
For non-quasi-analytic ultradierentiable functions and non-quasi-analytic
ultradistributions confer [5] and [6].
Denition 2.3. Let K R
n
be a compact set, M
p
satisfy (M.1) and
(NA). For f c
{M
p
},h
[K] we dene its norm by
(4) |f|
E
{M
p
},h
[K]
:= sup
xK,
[D

f(x)[
h
||
M
||
.
Let be an open set and K be a compact set. Topologies of ultradierentiable
classes are dened as follows.
c
{M
p
}
[K] = lim

h
c
{M
p
},h
[K],
c
{M
p
}
() = lim

K
c
{M
p
}
[K],
c
(M
p
)
[K] = lim

h0
c
{M
p
},h
[K],
c
(M
p
)
() = lim

K
c
(M
p
)
[K].
(5)
We dene c

as the strong dual of c

[K] and call it the set of ultradistributions

of the class supported by K. We also dene c

() :=
K
c

.
Let us dene the sheaf of ultradistributions. i
i
i
i
i
i
i
i
Structure of Ultradistributions 429
Denition 2.4. Let a sequence M
p
of positive numbers satisfy (M.0),
(M.1), (M.2)

, and
(6) limsup
p
p

p!
M
p
< .
We dene a presheaf F

on R
n
by
(7) F

() := c

(R
n
)/c

(R
n
),
where is any open set in R
n
and = (M
p
) or M
p
. We denote the corre-
sponding sheaf by T

. If M
p
satises (M.3)

, T

= T

. If M
p
satises (QA)
and (NA), then we call T

the sheaf of quasi-analytic ultradistributions of the

class .
Denition 2.5. For two classes and we dene their inclusion rela-
tions.
c

,
< c

.
(8)
Denition 2.6. A dierential operator P(D) :=

of innite
order is dened to belong to the class (M
p
) (resp. M
p
), if and only if there
exist such constants L and C (resp. for any L > 0 there exists such a con-
stant C) that [a

[ (CL
||
)/M
||
holds for any . We call this operator an
ultradierential operator of the class (M
p
) (resp. M
p
).
Denition 2.7. For a positive sequence M
p
satisfying (NA), dene its
associated function by
(9)

M(t) := sup
p
t
p
M
p
,
for t > 0.
3. Known Results
In this section, we review the known results on the structure theorems.
The structure theorem for distributions was proved by L.Schwartz.
Theorem 3.1 (cf. [7]). Any distribution f is locally represented as
(10) f = P(D)g,
where P(D) is a dierential operator of nite order with constant coecients
and g is a continuous function. i
i
i
i
i
i
i
i
430 Takashi Takiguchi
Extensions of this theorem for non-quasi-analytic ultradistributions and
for hyperfunctions are known. H. Komatsu [5] proved the structure theorem
for strongly non-quasi-analytic ultradistributions.
Theorem 3.2 (cf. [5]). Let the sequence M
p
satisfy the conditions, (M.1),
(M.2) and (M.3). f T

, where is (M
p
) or M
p
, is locally represented in
the form (10), where P(D) is an ultradierential operator of the class with
constant coecients and g is a continuous function.
This theorem was extended by R. W. Braun [1] for non-quasi-analytic
ultradistributions.
Theorem 3.3 (cf. [1]). Let the sequence M
p
satisfy the conditions, (M.1),
(M.2) and (M.3)

. For f T

, where is (M
p
) or M
p
, and for any class
satisfying < , there exist an ultradierential operator P(D) of the class
with constant coecients and an ultradierential function g of the class such
that the representation (10) locally holds.
In [3], A. Kaneko proved the structure theorem for hyperfunctions.
Theorem 3.4 (cf. [3]). Any hyperfunction f is locally represented as
(11) f = J(D)g,
where J(D) is a local operator with constant coecients, that is, J(D) is an
innite order dierential operator J(D) =

with the coecients sat-

isfying lim
||
||
_
[a

[! = 0, and g is an innitely dierentiable function.

The structure theorem for quasi-analytic ultradistributions is left open, to
prove which is our main purpose in this paper.
4. Fourier Transform of Non-analytic Functions and
Non-analytic Ultradistributions
In this section, we study the Fourier transform of non-analytic functions
and non-analytic ultradistributions. The properties proved in this section take
important roles to prove our main theorem. For a function f dened on R
n
,
we dene its Fourier-Laplace transform

f(), C
n
by
(12)

f() :=
_
R
n
e
ix
f(x)dx
when it is well-dened. i
i
i
i
i
i
i
i
Structure of Ultradistributions 431
Denition 4.1. Let M
p
be a sequence of positive numbers. A function
f c
(M
p
)
(R
n
) (resp. f c
{M
p
}
(R
n
)) belongs to o
(M
p
)
(resp. o
{M
p
}
) if
and only if for any k > 0 and h > 0 there exists a constant C = C
h,k
> 0
(resp. there exists a constant h > 0 and for any k > 0 there exists a constant
C = C
k
> 0) such that
(13) [D

f(x)[ C[h[
||
M
||
(1 +[x[)
k
,
for any multi-index . Let us dene
(14) o
{M
p
},h
= o
{M
p
}
; k, C, [D

(x)[ Ch
||
M
||
(1 +[x[)
k
.
For f o
{M
p
},h
, we dene its norm by
(15) |f|
S
{M
p
},h := sup
xR
n
,,k
[D

f(x)[
h
||
M
||
(1 +[x[)
k
.
For topologies of ultradierentiable classes, the following relations hold.
o
{M
p
}
= lim

h
o
{M
p
},h
,
o
(M
p
)
= lim

h0
o
{M
p
},h
.
(16)
The set o

is dened as the strong dual of o

, where = (M
p
) or M
p
.
Lemma 4.1 (cf. Proposition 3.2 in [5]). A sequence M
p
satises the con-
dition (M.1) if and only if
(17) M
p
= M
0
sup
t>0
t
p

M(t)
,
for t > 0.
Proposition 4.1. Assume that a sequence M
p
, p = 0, 1, 2, . . . , of pos-
itive numbers satises the conditions (M.0), (M.1), (M.2)

and (NA). Then

the following conditions are equivalent.
(i) The function

f is the Fourier-Laplace transform of f o
(M
p
)
(resp. f
o
{M
p
}
).
(ii) For any multi-index and h > 0 there exists a constant C = C
,h
(resp.
there exists a constant h > 0 and for any multi-index there exists a
constant C = C

> 0) such that i

i
i
i
i
i
i
i
432 Takashi Takiguchi
(18) [D

f()[
C

M(h[[)
, for R
n
.
Proof. Let us treat both (M
p
) and M
p
classes simultaneously.
(ii) (i); By virtue of Lemma 4.1, the following estimate holds.

x
f(x)

1
(2)
n
_
R
n
D

f()

)e
ix
d

1
(2)
n
_
||1

f()

d +
1

n
_
||>1
[[
n+1
(1 +[[)
n+1

f()

C
,1
M(h[[)
+ C
,2
_
||>1
1
(1 +[[)
n+1
d
min(||+n+1,||)

j=0
[[
||+n+1j
M(h[[)

C
,1
M(h[[)
+ C
,3
[[
||+n+1
M(h[[)
C

M
||+n+1
h
||+n+1
,
(19)
where C
,i
, i = 1, 2, 3, and C

are suitable constants. The condition (M.2)

yields that
(20)
M
||+n+1
h
||+n+1

GH
n+1
h
n+1
_
H
h
_
||
M
||
,
for some constants G and H. Therefore, (i) is obtained if the conditions on the
constants are properly interpreted according to the class (M
p
) or M
p
.
(i) (ii);
(21)

f() =
_
e
ix
_
D

(x

f(x))
_
dx.
Note that x

f(x) o

, = (M
p
) or M
p
. Then we have
(22)

f()

h
||
M
||
_
1
(1 +[x[)
(n+1)
dx.
Therefore
(23)

f()

inf
||
M
||
([[/h)
||

C

M([[/h)
,
which proves (ii) with an appropriate interpretation on the constants in accor-
dance with the class. i
i
i
i
i
i
i
i
Structure of Ultradistributions 433
Proposition 4.2. Assume that a sequence M
p
, p = 0, 1, 2, . . ., of pos-
itive numbers satises the conditions (M.0), (M.1), (M.2)

and (NA). If for

any h > 0 there exists such a constant C = C
h
(resp. there exist such constants
h > 0 and C > 0) that
(24) [f(x)[
C

M(h[x[)
,
then

f c
(M
p
)
(resp.

f c
{M
p
}
).
Proof.

f()

_
R
n
f(x)x

e
ix
dx

_
|x|1
[f(x)[[x[
||
dx + 2
n
_
|x|>1
[f(x)[[x[
||+n+1
(1 +[x[)
n+1
dx

_
_
|x|1
dx
_
sup
xR
n
C[x[
||

M(h[x[)
+
_
_
|x|>1
1
(1 +[x[)
n+1
dx
_
sup
xR
n
C[x[
||+n+1

M(h[x[)
C
1
M
||
[h[
||
+ C
2
M
||+n+1
[h[
||+n+1
,
(25)
where we have applied the fact that for a positive sequence M
p
, (M.1) is equiv-
alent to
(26) M
p
= sup
t>0
t
p

M(t)
.
By virtue of (M.2)

, there exist such constants C

3
and H that
(27) M
||+n+1
C
3
H
||
M
||
.
By (25) and (27), we have
(28)

f()

CH
||
max
_
1
h
||
,
1
h
||+n+1
_
M
||
,
which implies

f c
(M
p
)
(resp.

f c
{M
p
}
).
Theorem 4.1. (The Paley-Wiener theorem for non-analytic ultradis-
tributions) Let M
p
satisfy (M.0), (M.1), (M.2)

and (NA). For a compact

convex set K R
n
, the following conditions are equivalent. i
i
i
i
i
i
i
i
434 Takashi Takiguchi
(i)

f is the Fourier-Laplace transform of f c
(M
p
)
K

(resp. f c
{M
p
}
K

).
(ii) There exist such constants L > 0 and C > 0 (resp. for any L > 0, there
exists such a constant C > 0) that
[

f()[ C

M(L[[),
and

f() is an entire function in C
n
which satises that for any > 0
there exists such a constant C

that
(29) [

f()[ C

exp(H
K
(Im) + [[), C
n
,
where H
K
(y) := sup
xK
x y, y R
n
, is the supporting function of K.
(iii)

f() is an entire function in C
n
which satises that there exist such
constants L > 0 and C > 0 (resp. for any L > 0, there exists such a
constant C > 0) that
(30) [

f()[ C

M(L[[)e
H
K
(Im)
, C
n
.
This theorem seems to be known, however, it seems dicult to nd the
proof of this theorem in our form. Hence we shall give its proof.
Proof. (i) (iii); Since f c

(R
n
) c

(R
n
), there exist constants h
and C (resp. for any h > 0, there is a constant C) such that
(31) [ , f) [ C sup
xK,
D

(x)
h
||
M
||
, c

(R
n
).
Let
(32) (x) = exp(ix ), = + i C
n
.
Then there holds
(33) [

f()[ C sup
xK,
[[
||
h
||
M
||
[e
ix+x
[ Ce
H
K
(Im)

M
_
[[
h
_
.
(iii) (ii); Let any L > 0 be xed. For any > 0 there exists m N such that
(34) k! M
k
_

L
_
k
for k > m by virtue of (NA). Therefore
(35)

M(L[[) = sup
k
L
k
[[
k
M
k
C sup
k
([[)
k
k!
Ce
||
. i
i
i
i
i
i
i
i
Structure of Ultradistributions 435
(35) and the Paley-Wiener theorem for hyperfunctions (cf. [4]) give (ii).
(ii) (i); Let = (M
p
) or M
p
. We rst prove that f o

. By virtue of
Proposition 4.1 and the assumption ii), there holds for o

,
[, f)[ :=

_
(x)f(x)dx

=
1
(2)
n

_
,

f
_

1
(2)
n
_

()

f()

1
(2)
n
_
C
1

M(h[[)
C
2

M(L[[)d.
(36)
For = (M
p
), there exist some L, C
2
, and for any h > 0, there exists some C
1
such that (36) holds. Hence take h > 0 such that h > L then (36) converges.
For = M
p
, there exist some h, C
1
, and for any L > 0, there exists some C
2
such that (36) holds. Hence take L > 0 such that h > L then (36) converges.
The function f is then proved to be a linear map from o

to C. Take a sequence

n
in o

and replace in (36) by

n
, then the Lebesgue dominated
convergence theorem proves the continuity of f. Therefore, it is proved that
f o

.
The estimate (29) and the Paley-Wiener Theorem for hyperfunctions yield
that f is a hyperfunction with its support contained in K. The fact that the
space of the analytic functions /[K] is dense in c

[K] implies that f c

.
Therefore (i) is obtained.
The following proposition follows almost directly from Theorem 4.1.
Proposition 4.3. For the positive sequences M
p
and N
p
satisfying the
conditions (M.1) and (NA), the following conditions are equivalent.
(i) M
p
N
p
.
(ii) lim
p
(M
p
/N
p
)
1/p
= 0.
If M
p
and N
p
satises (M.2)

in addition, then above two condition are equiv-

alent to the following one.
(iii) M
p
< N
p
and (M
p
) < (N
p
).
By this proposition, we have M
p
N
p
for M
p
N
p
.
Denition 4.2. A function (t) > 0 dened for t > 0 is called sub-
ordinate if and only if it is continuous, monotonously increasing and (t)/t is i
i
i
i
i
i
i
i
436 Takashi Takiguchi
monotonously decreasing to zero as t , i.e.,
(37) lim
t
(t)
t
= 0.
Proposition 4.4 (cf. Lemma 3.10 in [5]). For positive sequences M
p
and
N
p
satisfying (M.1), the following conditions are equivalent.
(i) M
p
N
p
.
(ii) For any L > 0, there exists such a constant C > 0 that

N(t) C

M(Lt), for 0 < t < .

(iii) There exists such a subordinate function (t) that

N(t)

M((t)).
By virtue of Proposition 4.4, we obtain the following equivalent conditions.
Proposition 4.5. Let M
p
satisfy (M.1) and (NA). For a function f
dened on R
n
, the following conditions are equivalent.
(i) For any L > 0 there exists such a constant C > 0 that
(38) [f(x)[ C

M(L[x[), for x R
n
.
(ii) There exists such a subordinate function (t) that
(39) [f(x)[

M(([x[)), for x R
n
.
Proof. The proof of (ii) (i) is clear.
Let us prove (i) (ii). We dene (t), t > 0 by
(40) sup
|x|t
[f(x)[ =

M((t)).
By the denition, (39) holds. It is also trivial that (t) is monotonously in-
creasing. What is left to prove is that
(41) lim
t
(t)
t
= 0.
Let us assume the contrary to (41), that is, there exist a constant L > 0
and a sequence t
j
of positive numbers satisfying t
1
< t
2
< < t
j
< i
i
i
i
i
i
i
i
Structure of Ultradistributions 437
such that (t
j
) 2Lt
j
. Then for this constant L, there exists a constant C
such that
(42)

M(2Lt
j
)

M((t
j
)) = sup
|x|t
j
[f(x)[ C

M(Lt
j
).
Hence we obtain

M(2Lt
j
) C

M(Lt
j
), which contradicts to the fact that
(43) lim
t

M(t)
t
k
= ,
for any positive integer k.
5. Main Theorem
In this section, we prove our main theorem. As an preparation, we con-
struct the symbols of ultradierential operators in the non-analytic classes,
which serves as a key lemma to prove our main theorem.
Lemma 5.1. Let N
p
satisfy (M.0), (M.1), (M.2)

, (QA) and (NA).

For any subordinate function (t) there exist such a monotonously decreasing
positive sequence l
p
with lim
p
l
p
= 0 and such a constant A > 0 that
(44) [P()[ A

M(([[)),
where
(45) P() :=

p=0
(l
2p
[[)
2p
M
2p
.
Proof. If the subordinate function (t) satises lim
t
(t) < , then
the lemma is easily obtained, for example, by letting l
p
= 1/p. Therefore
what is left to prove is the case where lim
t
(t) = . Let us represent the
associate function

M by
(46)

M(t) = sup
p
t
p
M
p
= sup
p
p

q=1
t
m
q
,
where m
p
:=
M
p
M
p1
is an increasing sequence satisfying
(47) lim
p
m
p
= . i
i
i
i
i
i
i
i
438 Takashi Takiguchi
We dene the sequence

l
p
, p = 1, 2, . . ., by
(48)
_
m
p

l
p
_
= m
p
.
Monotonous increase of and m
p
together with (47) yields that the sequence
m
p
/

l
p
is monotonously increasing and
(49) lim
p
m
p

l
p
= .
Therefore the sequence
(50)

l
p
=
(m
p
/

l
p
)
m
p
/

l
p
is monotonously decreasing and satises lim
p

l
p
= 0.
If m
p
t < m
p+1
then
(51)
t
k
M
k
=
k

q=1
t
m
q
attains its supremum at k = p. Therefore for
(52) m
p
=
_
m
p

l
p
_
(t)
_
m
p+1

l
p+1
_
= m
p+1
,
there holds that
(53)

M((t)) =
p

q=1
(t)
m
q

q=1

l
q
t
m
q
=

l
1

l
p
t
p
M
p
=
(

l
p
t)
p
M
p
,
where

l
p
:=
p
_

l
1

l
p
and we have applied the estimate
(54) (t)

l
q
t,
for t m
p
/

l
p
and q p. Note that by virtue of (52), we obtain
m
p

l
p
t
m
p+1

l
p+1
.
By the denition,

l
p
> 0 is a decreasing sequence and satises lim
p

l
p
= 0.
For 0 <

l
p
t < 1,
(55)
(

l
p1
t)
p1
M
p1

l
p
t)
p
M
p
. i
i
i
i
i
i
i
i
Structure of Ultradistributions 439
For

l
p
t 1, (M.2)

implies
(56)
1
M
p1

GH
p
M
p
,
for some G and H, which yields
(57)
(

l
p1
t)
p1
M
p1

l
p
t)
p
M
p
GH
p
.
We have, by virtue of (55) and (57),
(58)
(

l
p
t)
p
M
p
G

p=0
(H

l
2p
t)
2p
M
2p
,
for any p = 0, 1, 2, , where H

:= max1, H. By (53) and (58), letting

l
p
:= H

l
p
yields
(59) A

M(([[)) P() := G

p=0
(l
2p
[[)
2p
M
2p
,
where G

:= max1, G, which proves the lemma.

Theorem 5.1. Let N
p
satisfy (M.0), (M.1), (M.2), (QA) and (NA).
Assume that f T

, where = (M
p
) or M
p
. Then for any class satisfying
< there exist g c

and an ultradierential operator P(D) of the class

such that the representation
(60) f = P(D)g,
locally holds.
Proof. It is sucient to prove that the representation (60) holds in some
neighborhood of the origin. By the denition, for any f T

there exists
f
1
c

such that f = f
1
in some neighborhood of the origin, where K R
n
is some compact set containing the origin in its inside. In the following proof,
it is capable of assuming that f c

without loss of generality.

I. The proof for (M
p
) class.
The ultradierential operator
(61)

P(D) :=

p=0
()
p
M
2p
, i
i
i
i
i
i
i
i
440 Takashi Takiguchi
belongs to (M
p
) class and satises that there exist such constants C
1
> 0 and
C
2
> 0 that
(62) [

P()[ C
1

M(C
2
[[),
for any R
n
. Theorem 4.1 yields that for f c
(M
p
)

there exist such

constants C and L that
(63) [

f()[ C

M(L[[), for R
n
.
Dene the ultradierential operator
(64) P(D) :=

p=0
(
L
C
2
)
p
M
2p
.
Then we have by (62) that there exists such a constant C

> 0 that
(65) [P()[ C

M(L[[),
for any R
n
. (63) and (65) yield that there exists such a constant C > 0
that
(66)

f()
P()
2

M(L[[)
.
for any R
n
. By Proposition 4.2, there holds
(67) g := T
1
_

f()
P()
2
_
c
{M
p
}
,
where T
1
is the inverse Fourier-Laplace transform operator. We have
(68) f(x) (P(D))
2
g(x).
By virtue of (M.2), we have (P(D))
2
is an ultradierential operator of the class
(M
p
). Therefore the theorem is proved for (M
p
) class.
II. The proof for M
p
class.
Let M
p
= (N
p
) or N
p
. L
p
:=
_
M
p
N
p
yields M
p
L
p
N
p
.
By Proposition 4.4, there exists such a subordinate function
1
that

L(t) =

M(
1
(t)), hence by Lemma 5.1, there exist such a positive decreasing sequence
l
(1)
p
with lim
p
l
(1)
p
= 0 and a constant A
1
> 0 that
(69) P
1
() :=

p=0
(l
(1)
2p
[[)
2p
M
2p
A
1

M(
1
([[)), i
i
i
i
i
i
i
i
Structure of Ultradistributions 441
for any R
n
. By the denition, P
1
(D) is an ultradierential operator of the
class M
p
. By virtue of Theorem 4.1 and Proposition 4.4, there exists such a
subordinate function
2
that
(70) [

f()[

M(
2
([[)),
for any R
n
. In view of Lemma 5.1, there exist such a positive decreasing
sequence l
(2)
p
satisfying lim
p
l
(2)
p
= 0 and a constant A
2
> 0 that
(71) P
2
() :=

p=0
(l
(2)
2p
[[)
2p
M
2p
A
2

M(
2
([[)),
for any R
n
. Therefore, we have
(72)

f()
P
1
()P
2
()

1
A
1
A
2

M(
1
([[))
=
1
A
1
A
2

L([[)
,
for any R
n
. Let us dene
(73) g := T
1
_

f()
P
1
()P
2
()
_
,
then Proposition 4.2 and (72) imply that g c
{L
p
}
c

. We have
(74) P
1
(D)P
2
(D)g(x) = f(x).
By (M.2), the ultradierential operatorP
1
(D)P
2
(D) belongs to the M
p
class.
Therefore the theorem is also proved for M
p
class.
As an application of the proof of Theorem 5.1, we obtain a modication
of Theorem 3.4.
Theorem 5.2. For any hyperfunction f and for any non-analytic class
, there exist such a local operator J(D) and g c

that the representation

(11) holds locally.
In order to prove this theorem, we modify A. Kanekos proof of Theorem
3.4 in [3] applying the fact that any ultradierential operator of non-analytic
classes is a local operator, that is, we take
(75) g := T
1
_

f()
J()P()
_
, i
i
i
i
i
i
i
i
442 Takashi Takiguchi
where J(D) is the local operator constructed in Lemma 1.2 in [3] and P(D) is
the ultradierential operator constructed in (64) for (M
p
) class (resp. in (71)
for M
p
class).
In the proofs of both Theorems 5.1 and 5.2, it is essentially important
to construct ultradierential operators of non-analytic classes (Lemma 5.1 and
(64)).
References
[1] R. W. Braun, An extension of Komatsus second structure theorem for ultradistributions,
J. Fac. Sci. Univ. Tokyo Sect. IA Math. 40 (1993), no. 2, 411417.
[2] L. Hormander, Between distributions and hyperfunctions, Asterisque No. 131 (1985),
89106.
[3] A. Kaneko, Representation of hyperfunctions by measures and some of its applications,
J. Fac. Sci. Univ. Tokyo Sect. IA Math. 19 (1972), 321352.
[4] , Introduction to hyperfunctions, Translated from the Japanese by Y. Yamamoto,
Kluwer Acad. Publ., Dordrecht, 1988.
[5] H. Komatsu, Ultradistributions. I. Structure theorems and a characterization, J. Fac.
Sci. Univ. Tokyo Sect. IA Math. 20 (1973), 25105.
[6] , Ultradistributions. II. The kernel theorem and ultradistributions with support
in a submanifold, J. Fac. Sci. Univ. Tokyo Sect. IA Math. 24 (1977), no. 3, 607628.
[7] L. Schwartz, Theorie des distributions, Publications de lInstitut de Mathematique
de lUniversite de Strasbourg, IX-X. Nouvelle edition, enti`erement corrigee, Hermann,
Paris, 1966.