Professional Documents
Culture Documents
HOLOMORPHIC EXTENSION
This content has been downloaded from IOPscience. Please scroll down to see the full text.
(http://iopscience.iop.org/0025-5734/61/2/A10)
View the table of contents for this issue, or go to the journal homepage for more
Download details:
IP Address: 150.164.25.130
This content was downloaded on 09/05/2017 at 13:39
On the structure of residue currents and functionals orthogonal to ideals in the space of
holomorphic functions
V M Trutnev and A K Tsikh
A. K. TSIKH
Introduction
In this article we investigate the properties of weakly holomorphic functions on analytic
sets which are complete intersections. Recall that a weakly holomorphic function on an
analytic set A is a function h which is defined and holomorphic at the regular points
of A and locally bounded on A. An analytic subset A of a complex n-dimensional
manifold X is said to be a complete intersection if it is purely (n — p)-dimensional and
is determined by the common zeros of ρ holomorphic functions. In the definition of
a complete intersection A it is sometimes required (see, for example, Abhyankar [1],
§12), that there exists a minimal system of defining functions for A, i.e., a system of
holomorphic functions / = (/1,. -., / p ) which at every point a € A = / - 1 ( 0 ) "divides"
any germ g € Oa which is equal to zero on the germ Aa of the analytic set A at the
point a; in other words, g \ Aa = 0 => g = qifi + • • • + qpfp for q3 € Oa. We shall show
(Proposition 1) that this algebraic requirement is equivalent to the following analytic
one: the set {z € A: {df\ Λ · · · f\dfp)z = 0} is nowhere dense in A, i.e., in a neighborhood
of every point a € A at least one of the Jacobians Jj = df/dzi and 2/ = ( 2 ^ , . . . , Zip) is
not identically equal to zero.
In §1 we prove (Theorem 1) that these Jacobians are universal denominators for com-
plete intersections, i.e., that every weakly holomorphic function on a complete inter-
section A can be represented locally as the restriction to A of a meromorphic function
with denominator Jj. This property of the Jacobians is then employed in §2 to obtain
a criterion for the holomorphic extendibility of weakly holomorphic functions (Theorem
3). More precisely, it is used to construct a current (hRf)° (for a weakly holomorphic
function h and a system / defining A as A = f~1{0)) with support on A such that this
1980 Mathematics Subject Classification (1985 Revision). Primary 32A10, 14M10, 32A25; Secondary
32C30, 32D15.
©1988 American Mathematical Society
0025-5734/88 $1.00 + $.25 per page
421
422 Α. Κ. TSIKH
π: Α Π U —• U" is an analytic cover. Consequently for all z" G U"\a, where σ is some
analytic subset in U", the preimage π~1(ζ") consists of the same number m of distinct
points z^v\z"), ν = l,...,m, which depend holomorphically on z" 6 U"\a. For the
functions /, defining A we use the expansions given by Hefer's theorem ([7], Paragraph
50, Theorem 2)
9=1
Let Η(ζ',ς',ζ") denote the determinant of the matrix formed by the coefficients Pjq of
these expansions. We consider the function
h(z) =
{ζ € Α: df = 0} (here and in what follows df denotes the exterior product dfi Λ · · · Λ dfp)
has nonempty interior in A, i.e., that it consists of several irreducible components of
A. As in the proof of Theorem 1, after applying a nonsingular linear transformation
we select a coordinate system ζχ,...,ζη and a neighborhood U — U' χ U" of the point
0 £ C " so that the projection π: Α Π U —> U" is an analytic cover. By our hypothesis
the Jacobian J = df/dz' is identically equal to zero on some irreducible components of
A, the union of which we denote by Αλ; we denote the union of the remaining irreducible
components by yl2. Let g be a function, holomorphic in U, which is equal to zero at
some point α Ε Αχ (if Αι = 0 we set g = 1). By construction the function gj is equal
to zero on A, i.e., it belongs to the ideal I A- At the same time, there does not exist for
gj a representation of the form
since the regularity of the sequence fi(z', α " ) , . . . , fp(z', a") at the point a' implies (see
[9], pp. 650-651, and [10], Theorem 4.5) that a representation of the form
is possible only when g(a) = g(a',a") = 0, while g(a) φ 0 by our construction. We obtain
a contradiction to the minimality of the defining system /.
We now assume that the set {z e A: df = 0} is nowhere dense in A. We select
a coordinate system zi,..., zn so that the Jacobian J = df/dz' is different from zero
almost everywhere on A and, moreover, that the projection ττ: AC\U —> U" is an analytic
cover. Therefore the system of functions f{z',0") = (/i(*',0"),... ,/ p (z',0")) has z' = 0'
as an isolated common zero. We denote by /o(/) and 70<{f{z', 0")) the ideals in the rings
Oo and Ο$> generated by the corresponding systems f(z) and f(z', 0"). We need to prove
that if g belongs to the ideal I A (i-e., if g\ A = 0 ) then g belongs to 7o(/). We select
a collection of monomials ei(z'),..., eM(z') that constitute a basis for the local algebra
Oo'/Io'{f(z',0")). By the well-known division theorem ([11], Chapter 2) the germ g can
be represented in the form
) + --- + cll(z")eli(z') + <p, <pelo(f), (1.2)
where Cj(z") 6 Oo" are defined uniquely by the germ g (see [12]). Since for every
z" £ U"\a (recall that σ is the discriminant set for the projection π: Α Π U —> U") the
Jacobian J is different from zero at the points z^u\z") € π~1(ζ") for i> = 1,... ,m (in
fact m = μ; see [12]), the roots z'^{z") of the system of equations f(z', z") = 0, where
z" is fixed and lies in U"\a, are simple. Therefore from the fact that g \ A — 0 it follows
that g belongs to the ideals
By Cartan's Theorem A ([8], p. 243) for fixed z" € U"\a there are q,{z',z") e O{U')
such that g = q\f\ + • • •-\-qPfP, i.e., in the expansion (1.2) for all z" € U"\a the functions
Cj(z") = 0. By the uniqueness theorem, Cj = 0 and g Ε /ο(/)· Thus we have proved the
minimality of the system /.
We prove finally that if the system / is minimal for the analytic set A = /~ x (0),
then S :— {/ = df = 0} coincides with the set of singular points sng A The inclusion
sngA C S follows from the implicit function theorem; therefore it suffices to establish
that reg.4 C A\S. Let us assume the contrary, that at some regular point α € A the
exterior derivative {df)a = 0. We "straighten" the analytic set A in a neighborhood
of a, i.e., we perform a biholomorphic change of coordinates ζ —<· w which sends the
point z = a t o » = 0 s o that a portion of A in a neighborhood of a takes the form
WEAKLY HOLOMORPHIC FUNCTIONS ON COMPLETE INTERSECTIONS 425
A* = (/*)- χ (0) = {wi = • • • = wp = 0}, where f*{w) = f{z{w)). After this change of
variables the exterior products (df)z and (df*)w(z) vanish simultaneously; therefore by
our hypothesis {df*)o = 0. At the same time, by the minimality of /, as we have shown
above, the set S Π Ua passes to the set S* which is nowhere dense in A*. Consequently
there is a sequence of points {u/fc'} C A* converging to 0 such that (df*)w(,k) φ 0. Since
the points w^ have the form (0'w"^), this means that in some ball β 0 ' C C p centered
at 0' the system of equations f*{w', w"^) = 0 has the unique root w' = 0 in B0' (since
(/*)~x(0) = {w' = 0}) of multiplicity 1 for every fixed value of the parameter w"^k\
while the system f*(w',0") = 0 has the multiple root w' = 0 for the limiting value
0" = limu>"'fe'. Thus we have arrived at the contradiction with the Rouche principle.
This completes the proof.
§2. The extension of weakly holomorphic functions from analytic sets
Here we consider the problem of the holomorphic extendibility of weakly holomorphic
functions from analytic subsets which are complete intersections. We shall start from
the well-known scheme for the solution of this problem, which in its clearest form was
described by Chirka ([13], p. 144). For analytic sets of codimension 1 this consists in the
following. Let an analytic set Λ in a manifold X be denned globally by a single equation
/ = 0 for / 6 O(X), where / is a minimal denning function for A. We consider the
current [1//], which is an extension to the space D2n(X) (by the Hahn-Banach theorem)
of the integration functional
2n
-LT
defined on those forms φ € D (X) all of whose partial derivatives are equal to zero on A.
Let us assume that a weakly holomorphic function h on A extends to a smooth function
h! on X, so that the current h'd[l/f] is defined. The current d[l/f] is concentrated
on A, and by the minimality of / is a current of measure type; therefore for any other
smooth extension h" we have the equality h"d[l/f] = h'd[l/f]. It follows that each
holomorphic extension h" of h satisfies d[h"f] = h'd[l/f], and moreover h admits a
holomorphic extension if and only if the current h'd[l/f] is d-exact (see [13], p. 145).
Thus any holomorphic extension of h has the form ψ/, where φ is a solution of the
equation Θφ = h'd[l/f] and h' is a smooth extension of h.
This scheme becomes more efficient if we observe that the currents [1//] and d[l/f]
can be realized as the following limits:
The existence of these limits was proved by Herrera and Liebermann [14]; they are called
the principal value current and the residual current respectively. As a consequence of
these realizations we give an "intrinsic" definition of the current hd[l/f] (i.e., a definition
in which a smooth extension of h is not required) and formulate a criterion for the
extendibility of weakly holomorphic functions. We shall carry out the arguments for
analytic sets of arbitrary codimension; therefore we give the generalizations of these
currents in codimension ρ > 1, which were considered by Coleff and Herrera [15].
Let V be an analytic set of pure dimension η in a manifold X (dimX > n), and
let / = (/i,..., fp): V —• C p be a holomorphic mapping, where the components fj are
holomorphic in the ambient manifold X, with ρ < η. We shall be interested only in the
case when / is a complete intersection, i.e., when /~ x (0) is an analytic subset in V of
pure dimension η — p. Associated to the mapping / is the following pair of semianalytic
426 Α. Κ. TSIKH
chains ("tubes"):
The notation RfPfp is explained by the fact that the "tube" De(f) is the intersection
of the "tube" Te>{f) (here / ' = (/Ί,... ,/ p _i)) with the set {|/p| > ε ρ }; therefore the
functional (2.2) is a residual current relative to /' and a principal value relative to the
latter function fp.
We note the following properties of the currents (2.1) and (2.2) (see [15]):
(1) The currents (2.1) and (2.2) are currents of bidegree (n, n — p) and (η, η - ρ + 1)
with supports in A — f~1(0) and A' = /' - 1 (0) respectively.
(2) d(Rf,Pfp) — (—l)p~1Rf, which implies that the residual current Rf is <S-closed.
(3) The restrictions of the currents (2.1) and (2.2) to any compact subset Κ C V have
finite orders, i.e., these currents extend continuously to the class of Ν times continuously
differentiable forms on X with supports in a neighborhood of K, where Ν is any natural
number.
There is an important relation between the residual current and the principal value
which is different from that described in property (2). This relation arises from the
natural desire to calculate the residual current by fibering the set V into p-dimensional
fibers Bt, each of which intersects the subset A = /~ 1 (0) in discrete points, and then
taking the residues of the form ω//χ · • · fp | a e at the points of the intersections AC\Bt.
We shall demonstrate this by deriving, using a method borrowed from [6], §16.3, Theorem
1, the well-known Poincare-Lelong formula
[ n-l,n-
7Λ=/"1(0)
To conform to our notation, by property (2) this formula can be written in the form
REMARK I . The form in the integrand in (2.6) is the multiple residue-form ω/df \ A —
ω I df\ A • • • Λ dfp \ A of the semimeromorphic form ω/f = ω/fi • • • fp ([16], §2); therefo
(2.6) can be written as i?/(w) = (2ni)p(Pj)A(Ju/df). In the case when A has no
singularities, the residue-form is smooth on A; consequently, we may omit the limit in
(2.6) and arrive at the classical version of the Leray formula:
where η — Α Π suppw, and δεη is the multiple coboundary of the compact chain 7
which coincides with the "tube" T£(f). For singular sets A the residue-form ω/df \ A
has singularities at the singular points of A; therefore to preserve (2.7) we are forced to
replace the integral over 7 on the right by its principal value.
We return to the problem of extending weakly holomorphic functions from analytic
sets. Let A be an analytic set in a domain G C C " which is a complete intersection, and
let / = (fi,..., fp) be a set of minimal denning functions for A. If a weakly holomorphic
function h on A admits a holomorphic (or class C°°) extension to G, then by Theorem
2 we have
(ΛΛ/)(ω) := Λ,(Λω) = {2πi)p{PJ)A(Jhω/df) (2.8)
for ω Ε Dn'n~p(G). Thus to define the current hRf we need only the values of the
function h on A, and the right side of (2.8) can serve as an "intrinsic" definition of the
current hRf, where as h we may take any weakly holomorphic function on A. Thus under
the hypothesis that the Jacobian J = df/dz1 is not equal to zero almost everywhere on
A, the "intrinsically" denned current hRf, which we denote by (hRf)°, is given by
im f
= (2π*) ρ lim
6
°
where the right side is the value at the form \ιω Ε Dn'n~p of the current which is equal
to the principal value on A relative to the function J 2 .
We now formulate a criterion for the extendibility of weakly holomorphic functions.
THEOREM 3. Let A — {ζ Ε G: f\{z) — • • • — fp(z) = 0} be a complete intersection
in a domain G C C " , where f = (f\,..., fp) is a minimal defining system for A. Then
a weakly holomorphic function h on A extends locally holomorphically from A if and only
if the current (hR/)° defined by (2.9) is d-closed.
PROOF. Let a E A and let h be a weakly holomorphic function on A. If h extends in
some neighborhood Ua C G to a holomorphic function h', then according to the definition
(2.9) and formula (2.8) we have (hRf)° = h'Rf, and since Rj is a d-closed current (see
property (2)) we have
d(hRf)° = B(h'Rf) = dti ARf + tidRj = 0.
WEAKLY HOLOMORPHIC FUNCTIONS ON COMPLETE INTERSECTIONS 429
Now let us assume that 3(hRf)° = 0. This means that {hRf)°(8<p) = 0 for any form
φ G Dn'n~p~x(G). If Ua is a neighborhood of the point α € A in which the extension h
of the function hi is holomorphic (according to Theorem 1), then by (2.10) we have
(recall that ζ' = (ζχ,... ,zp)). The set of integration in (2.11) lies in the regular part
of A, and on it the integrand form is the differential d\(h/J)(<p/df)] (here we are using
the fact that the degree of the form φ/df \ A = φ/Jdz' \ A relative to the holomorphic
variables ζ is equal to η — ρ, i.e., the dimension of A). In this case by Stokes' theorem
we obtain from (2.11) that
(3-1)
where for brevity we denote by z'"' the preimages of the point f(z) with the indices
chosen so that z^ = z. The function Β(ς, ζ) is holomorphic on the product Π χ Π, and
at each point ζ € Π at which the Jacobian J = df/dz of the mapping / is different
from zero we have the integral formula of Bishop for a meromorphic function h/J with
A € Ο(Π)
for the Sj sufficiently small (here we have the global extendibility of hJ, since Π is a
domain of holomorphy). Then by Bishop's formula
Κ{ς)Β{ς,ζ)άς
J(Z)
where Τ is the distinguished boundary of the polyhedron Π. We note that Jd<; =
dfi Λ · · · Λ dfp Λ dgi Λ · · • Λ dgn-p, and for ζ € A we have fj(z) = 0 for j = 1 , . . . ,p.
Further, Τ is the multiple coboundary <5ΡΓ of the cycle Γ; therefore, for ζ € Δ, (3.3)
gives
hJB($, z)df A dg
= 1 f
(2*i)»B(z,z)Js,rJlfl>fj
Now by Leray's formula we obtain (3.2).
Starting from the integral representation of Weil ([7], Paragraph 30, Theorem 1),
we can obtain a representation for weakly holomorphic functions which is analogous to
(3.2). As before, let A = / - 1 ( 0 ) , where the Jacobian J = df/dz' relative to the variables
z' — (zi,..., Zp) is not identically equal to zero on each irreducible component of A. We
assume that the polyhedron Δ = {ζ € A: \ZJ\ < rj, j — p+ l , . . . , n } is relatively
compact in A. Let us denote by Η the determinant of the matrix composed from the
Hefer expansions of the functions fi,...,fp, zp+i, • • •, zn (see (1.1)). In this case for any
weakly holomorphic function h on Δ we have the integral representation
... 1 f h{$)Hh,z)df ., . , ,
hiz) GA sngA (3 4)
-(2mr-Pj{z)Jr V-V ' * \ ' ·
where Γ is the distinguished boundary of the polyhedron Δ containing no critical points
of A a n d ς" - ζ" = {ςρ+1 - z p + 1 ) · • • ( ? „ - z n ) .
We find an integral representation for the extending function. Using the notation
employed in (3.4), we have
THEOREM 5. A weakly holomorphic function h in a closed polyhedron A C A extends
holomorphically to a full-dimensional neighborhood of A if and only if
In other words, for strongly holomorphic functions in the integral representation (3.4) the
Jacobian J can be carried under the integral sign; as a result we arrive at the integral (3.5),
•which is holomorphic in ζ in some full-dimensional neighborhood of the polyhedron Δ
and which therefore gives a holomorphic extension of the strongly holomorphic function h
using only the values of this function on A, more precisely, on the distinguished boundary
of the polyhedron Δ.
PROOF. Let h be extended to a holomorphic function h! in some neighborhood U of
Δ. Then by the Weil formula we have
Δ ( 3
·6)
z €
z")' '
where-/= {ς e U: |/,-(?)| =£j,j = l , . . . , p , |ft| = rk, k = p+ 1,... ,n}, and the Sj are
sufficiently small that η C U. Next we note that 7 is the multiple coboundary of the
distinguished boundary Γ; therefore, since άς' Λ άς" = dfi Λ · • · Λ dfp Λ άς"/J, we obtain
(3.5) from (3.6) using Leray's formula.
Conversely, if a weakly holomorphic function h is represented as the integral (3.5),
then the integral, which is holomorphic in some neighborhood of Δ, is precisely the
extending function for h.
REMARK 2. Using (3.5), we can extract information about the extending function
from properties of the function h defined on A. For example, if A = {/(21,2a) = 0}
is an algebraic curve in C 2 and h is a strongly holomorphic function on A which has
polynomial growth at infinity (as ζ \ A —> 00), then h admits a polynomial extension in
C 2 . In order to verify this, we expand the polynomial Η{ς,ζ) as a sum ^icjW)ej{z) °f
polynomials with separated variables. Further, we note that for any fixed ζ €E C 2 there
is a constant r° so that for all r > r° the cycles Tr = {ς e Α: \ς%\ = r} are mutually
homologous in the domain of regularity of the integrand form in (3.5) (it suffices to
choose r° so that all singular points of A fall inside the contour Tro and the point ζ
in the domain Gro = {z e C 2 : |-2r21 < r0}. Since the functions hcj/J have polynomial
growth as ς | A —• 00, then by expanding the fraction (f2 — ^ 2 ) - 1 as a geometric series
for (ς, ζ) e Γ Γ χ Gr we obtain that for r° » 1 the integrals
fh(zi,z2)
where k is the maximum number in the m-tuple {fcj,..., km}.
THEOREM 6. Any form ω of type (4.2) outside of its pole set A = {/ = 0} is
cohomologous to a rational differential form with a pole of the first order of A, i.e., a
form of type ω' = hdz\ Λ dz2/f, where h is an entire function.
PROOF. Let A» = j4\sngA denote the set of regular points of the complex algebraic
curve A. We choose a coordinate system in C 2 so that the partial derivative f'Z2 is not
identically equal to zero on each irreducible component of A. Relative to the manifold
X : = C 2 \sng.4 and its submanifold A» (i.e., relative to the pair (X,A*)) we consider
the exact cohomology sequence of Leray ([16], §32):
X = {hdz1lf'X2)\A., (4-5)
where h is an entire function.
For the proof of this lemma we need
LEMMA 2. Let 9t be a one-dimensional complex manifold obtained from a compact
Riemann surface by deleting a nonempty finite set of points, and let {bj} denote a fi-
nite collection of points on 91 Then the cohomology group H1(9\\\Jbj) is generated by
meromorphic l-forms with poles of the first order at the points bj.
PROOF OF LEMMA 2. Let φ be an arbitrary closed differential 1-form on
and let r3 denote its residues at the points bj. It is well known (see, for example, [9], p.
233) that the Mittag-Leffler theorem holds for forms on 9Ί: there exists a meromorphic
differential 1-form a with poles of the first order at the points bj and residues r3 at these
points. This means that in the Leray cohomology sequence for the pair (iK,\Jb3),
the images δ*[φ\ and δ*[α] coincide; and by the exactness of this sequence the form φ is
cohomologous to a form of type α + αχ, where αχ is a differential 1-form which is regular
and closed on 9Ί. Using the fact that UH is a Stein manifold (9t is a noncompact Riemann
surface), we replace αχ by a holomorphic 1-form which is cohomologous to it on iK. As
434 Α. Κ. TSIKH
L λ χ = Ο, λ e 0{Ua),
holds, where Aa is the germ of the curve A at the point ο and δ is sufficiently small
(4.7)
(in order to follow Theorem 3 precisely, we would need in (4.7) to take the limit of the
integral as δ —* 0 and take smooth functions for λ; however, it is easy to verify that we
need only let λ be holomorphic). It is clear that the set of integration in (4.7) coincides
with the intersection ηα Π A», where ηα is the boundary of the polyhedral neighborhood
{z € Ua: I/I < ε, \f'Z2\ < δ} of the point a, where the intersection ηα Π A* consists of ra
closed real curves Cj for j = 1,... ,ra (ra is the number of local irreducible components
of A at the point a). The preimage π"1 (a) consists of ra points bi,...,bTa, and K~1(CJ)
is a circle on the normalization UH surrounding bj. Since π*χ has poles of the first order
at the points bj and (n*X)(bj) = λ(ο) for all j from 1 to r a , we have
β8(π·χ). (4.8)
WEAKLY HOLOMORPHIC FUNCTIONS ON COMPLETE INTERSECTIONS 435
In the particular case when λ Ξ 1 from (4.4) by the exactness of the sequence (4.3) we
have
This together with (4.8) verifies the validity of (4.7). We conclude that h extends to an
entire function, which concludes the proof of Lemma 2.
Returning to the proof of the theorem, we note that
(hdZl/f'Z2) Λ (df/f) = hdz! Λ dz2/f =: ω'.
This means that δ*[ω ] = χ where χ is the form (4.5), i.e., δ*[ω] = δ*[ω1]. But since
1
H2(X) = Η2 (C2\sng A) ~ 0 by the exact sequence (4.3), we thus obtain the equality
[ω] = [ω']. This proves the theorem.
REMARK 4. It is not known to the author whether the form ω1 in Theorem 6 can
be chosen to be rational, i.e., whether Λ is a polynomial. One could seek the answer
to this question in the following way. Recall that the entire function in (4.5) is none
other than the extension of the weakly holomorphic function h from (4.6). According to
Remark 2, this extension can be chosen to be polynomial if the form (4.6) has a polar
singularity at infinity. But the form (4.6) is of type χ = (π~1)*<ρ, where ττ: 9Ί —• A
is a normalization of the curve A. Moreover, *K is obtained from 9t (the normalization
of the closure A C P2) by the deletion of points corresponding to points at infinity on
A. Therefore our question reduces to the following: is it true that the one-dimensional
cohomology of a punctured compact Riemann surface can be realized by meromorphic
1-forms with poles at the deleted points? According to the algebraic de Rham theorem
(see [4]) the answer to this question is positive if the punctured compact Riemann surface
is an affine curve.
Kirenskii Institute of Physics
Siberian Branch, Academy of Sciences of the USSR Received 6/ JUNE/86
Krasnoyarsk
BIBLIOGRAPHY
1. Shreeram Abhyankar, Concepts of order and rank on a complex space, and a condition for normality,
Math. Ann. 141 (1960), 171-192.
2. Edgar Lee Stout, An integral formula for holomorphic functions on strictly pseudoconvex hypersurfaces,
Duke Math. J. 42 (1975), 347-356.
3. Gennadi M. Henkin and Jiirgen Leiterer, Global integral formulas for solving the d-equation on Stein
manifolds, Ann. Polon. Math. 39 (1981), 93-116.
4. M. F. Atiyah, R. Bott, and L. Garding, Lacunas for hyperbolic differential operators with constant
coefficients. II, Acta Math. 131 (1973), 145-206.
5. M. Hervi, Several complex variables. Local theory, Tata Inst. Fund. Res., Bombay, and Oxford Univ.
Press, London, 1963.
6. Ε. Μ. Chirka, Complex analytic sets, "Nauka", Moscow, 1985. (Russian)
7. B. V. Shabat, Introduction to complex analysis. Part 2, 3rd ed., "Nauka", Moscow, 1985.(Russian)
8. Robert C. Gunning and Hugo Rossi, Analytic functions of several complex variables, Prentice-Hall,
Englewood Cliffs, N. J., 1965.
9. Phillip Griffiths and Joseph Harris, Principles of algebraic geometry, Wiley, 1978.
10. L. A. Aizenberg and A. P. Yuzhakov, Integral representations and residues in multidimensional
complex analysis, "Nauka", Novosibirsk, 1979; English transl., Amer. Math. Soc, Providence, R.I., 1983.
11. V. P. Palamodov, Linear differential operators with constant coefficients, "Nauka", Moscow, 1967;
English transl., Springer-Verlag, 1970.
12. , On the multiplicity of a holomorphic mapping, Funktsional. Anal, i Prilozhen. 1 (1967), no.
3, 54-65; English transl. in Functional Anal. Appl. 1 (1967).
13. Ε. Μ. Chirka, Currents and their applications, Appendix to the Russian transl. of Reese Harvey,
Holomorphic chains and their boundaries, "Mir", Moscow, 1979, pp. 122-154.
436 Α. Κ. TSIKH
14. Μ. Herrera and D. Lieberman, Residues and principal values on complex spaces, Math. Ann. 194
(1971), 259-294.
15. Nicolas R. Coleff and Miguel G. Herrera, Les courants re'siduels associes a une form meromorphe,
Lecture Notes in Math., vol. 633, Springer-Verlag, 1978.
16. Jean Leray, Le calcul differentiel et integral sur une variete analytique complexe (Probleme de Cauchy,
HI), Bull. Soc. Math. France 8 7 (1959), 81-180.
17. M. Pettersson, Residues, currents, and their relation to ideals of holomorphic functions, Report, Dept.
Math., Univ. Uppsala, Uppsala, 1984.
18. Karlheinz Spallek, Differenzierbare und holomorphe Funktionen auf analytischen Mengen, Math. Ann.
161 (1965), 143-162.
19. I. R. Shafarevich, Basic algebraic geometry, "Nauka", Moscow, 1972; English transl., Springer-
Verlag, 1974.
20. Errett Bishop, Mappings of partially analytic spaces, Amer. J. Math. 83 (1961), 209-242.
Translated by J. NUNEMACHER