APP下载

Second Main Theorem for Meromorphic Maps into Algebraic Varieties Intersecting Moving Hypersurfaces Targets∗

2021-11-03LibingXIETingbinCAO

Libing XIE Tingbin CAO

Abstract Since the great work on holomorphic curves into algebraic varieties intersecting hypersurfaces in general position established by Ru in 2009, recently there has been some developments on the second main theorem into algebraic varieties intersecting moving hypersurfaces targets. The main purpose of this paper is to give some interesting improvements of Ru’s second main theorem for moving hypersurfaces targets located in subgeneral position with index.

Keywords Algebraic varieties, Holomorphic curves, Nevanlinna theory, Moving hypersurfaces

1 Introduction and Main Results

It is well-known that in 1933, Cartan generalized Nevanlinna theory for meromorphic functions to the case of linearly nondegenerate holomorphic curves into complex projective spaces intersecting hyperplanes in general position, and conjectured that it is still true for moving hyperplanes targets. From then on,higher dimensional Nevanlinna theory has been studied widely(see[7,11,16]). In 2009,Ru[13]proposed a great work on second main theorem of algebraically nondegenerate holomorphic curves into smooth complex varieties intersecting hypersurfaces in general position, which is a generalization of the Cartan’s second main theorem and his own former result(see [12])completely solving the Shiffman’s conjecture(see [14])corresponding to the Corvaja-Zannier’s theorem (see [1]) in Diophantine approximation (see [4]).

Thus,it is natural and interesting to investigate the Ru’s second main theorem into complex projective spaces and even into complex algebraic varieties for the moving hypersurfaces targets.Based on their affirmation of the Shiffman’s conjecture for moving hypersurfaces targets (see[2]),recently,Dethloffand Tan[3]continue to prove successfully the following special case where the coefficients of the polynomialsQj’s are constant and the varietyVis smooth, namely, the Ru’s second main theorem (see [13]).

Theorem 1.1(see [3])Let V⊂Pn(C)be an irreducible(possibly singular)variety of dimension ℓ, and let f be a nonconstant holomorphic map ofCinto V. LetD ={D1,···,Dq}be a family of slowly moving hypersurfaces(with respect to f)in general position, and letQ = {Q1,···,Qq}be the set of the defined homogeneous polynomials ofDwithdegQj=dj(j=1,···,q)and Qj(f)/≡0for j=1,···,q. Assume that f is algebraically nondegenerate overKQ. Then, for any ε>0,

holds for all r outside a set with finite Lebesgue measure.

For the special caseV= Pn(C),Quang [8] recently gave a second main theorem with truncated counting functions for meromorphic mappings into Pn(C) intersecting a family of moving hypersurfaces in subgeneral position, which any be possibly good at the uniqueness problems of meromorphic mappings.

Theorem 1.2(see [8])Let f be a nonconstant meromorphic map ofCmintoPn(C). Letbe a collection of slowly moving hypersurfaces in N-subgeneral position withdegQj=dj(1 ≤j≤q). Assume that f is algebraically nondegenerate overKQ. Then, for any ε>0,

holds for all r outside a set with finite Lebesgue measure, where

with

where d:=lcm(d1,···,dq)is the least common multiple of all{di}, and

In this paper, we mainly combine the methods in [3, 8, 17] together and adopt the new concept of the index of subgeneral position due to Ji-Yan-Yu [6] to obtain some interesting developments of Ru’s second main theorem for moving hypersurfaces targets, which are improvements and extensions of Theorems 1.1–1.2.

According to [6], we can give a similar definition for moving hypersurfaces located inmsubgeneral position with indexk.

Definition 1.1Let V be an algebraic subvariety ofPn(C). Let{D1,···,Dq}be a family of moving hypersurfaces ofPn(C). Let N and κ be two positive integers such that N≥dimV≥κ.

(a)The hypersurfaces{D1,···,Dq}are said to be in general position(or say in weakly general position)in V if there exists z∈Cm(if this condition is satisfied for one z∈Cm,it is also satisfied for all z except for an analytic set of codimension≥2)for any subset I⊂{1,···,q}with ♯I≤dimV+1,

(b)The hypersurfaces{D1,···,Dq}are said to be in N-subgeneral position in V if there exists z∈Cm(if this condition is satisfied for one z∈Cm, it is also satisfied for all z except for an analytic set of codimension≥2)for any subset I⊂{1,···,q}with ♯I≤N+1,

(c)The hypersurfaces{D1,···,Dq}are said to be in N-general position with index κ in V if D1,···,Dqare in N-subgeneral position and if there exists z∈Cm(if this condition is satisfied for one z∈Cm, it is also satisfied for all z except for an analytic set of codimension≥2)for any subset I⊂{1,···,q}with ♯I≤κ,

(Here we setdim ∅=−∞).

Now, we state our main result which is an improvement and extension of the above two theorems concerning moving hypersurfaces targets located in subgeneral position with index.Theorem 1.1 is just the following result for the special case wheneverN=dimVandκ=1.

Theorem 1.3Let f: Cm→V⊂Pn(C)be a nonconstant meromorphic map, where V is an irreducible algebraic subvariety of dimension ℓ. LetQ = {Q1,···,Qq}be a collection of slowly moving hypersurfaces in N-subgeneral position with index κ in V, anddegQj=dj(j=1,···,q). Assume that f: Cm→V is algebraically nondegenerate overKQ. Then, for any ε>0,

holds for all r outside a set with finite Lebesgue measure.

WhenV=Pn(C),we can have the following second main theorem with truncation, and thus Theorem 1.2 is just the special case wheneverκ=1.

Theorem 1.4Let f be a nonconstant meromorphic map ofCmintoPn(C). Let{Qi}qi=1bea collection of slowly moving hypersurfaces in N-subgeneral position with index κ,anddegQj=dj(1 ≤j≤q). Assume that f is algebraically nondegenerate overKQ. Then, for any ε>0,

holds for all r outside a set with finite Lebesgue measure, where

with

where d:=lcm(d1,···,dq)is the least common multiple of all{di}, and

Remark that very recently, Yan and Yu [17] considered the nonconstant holomorphic curve from C into Pn(C) instead of algebraically nondegenerate and improved Theorem 1.2 without truncation. Thus it is interesting to ask the following question.

Question 1.1In Theorem 1.3 or Theorem 1.4, is it possible to obtain a second main theorem if the condition “fis algebraically nondegenerate over KQ” is omitted?

The remainder is the organization as follows. In the next section, we introduce some basic notions and auxiliary results from Nevanlinna theory. Sections 3–4 are the proofs of Theorems 1.3–1.4, respecitively, in which the methods to deal with moving targets by Dethloff-Tan [3],Yan-Yu [17], and the techniques to deal with hypersurfaces in subgeneral position instead of Nochka’s weights owing to Quang [8–9] are used in this paper.

2 Basic Notions and Auxiliary Results from Nevanlinna Theory

2.1 The characteristic function in Nevanlinna theory

We set ‖z‖=(‖z1‖2+···+‖zm‖2)12forz=(z1,···,zm)∈Cmand define

Define

LetFbe a nonzero meromorphic function on a domain Ω in Cm.For a setα=(α1,···,αm)of nonnegative integers, we set |α|=α1+···+αmand

We denote byν0F,ν∞FandνFthe zero divisor,the pole divisor,and the divisor of the meromorphic functionF, respectively.

For a divisorνon Cmand for a positive integerMorM=∞,we set

The counting function ofνis defined by

Similarly, we defineN(r,ν[M]) and denote it byN[M](r,ν).

Letϕbe a nonzero meromorphic function on Cm. Define

For brevity we will omit the character [M] ifM=∞.

Letf: Cm→Pn(C) be a meromorphic mapping. For arbitrarily fixed homogeneous coordinates (w0:···:wn) on Pn(C),we take a reduced representation=(f0,···,fn),which means that eachfiis a holomorphic function on Cmandf(z)=(f0(z),···,fn(z)) outside the analytic setI(f)={f0=···=fn=0} of codimension ≥2.SetThe characteristic function offis defined by

Letϕbe a nonzero meromorphic function on Cm,which is occasionally regarded as a meromorphic map into P1(C).The proximity function ofϕis defined by

The Nevanlinna’s characteristic function ofϕis defined as follows

Then

The functionϕis said to be small (with respect tof) if ‖Tϕ(r) =o(Tf(r)),here the notion ‖means that the property holds possibly outside a set with finite Lebesgue measure.

We denote by M (resp. Kf)the field of all meromorphic functions(resp. small meromorphic functions with respect tof) on Cm.

2.2 Family of moving hypersurfaces and the first main theorem

Denote by HCmthe ring of all holomorphic functions on Cm. LetQbe a homogeneous polynomial in HCm[x0,···,xn] of degreed≥1.Denote byQ(z) the homogeneous polynomial over C obtained by substituting a specific pointz∈Cminto the coefficients ofQ.We also call a moving hypersurface in Pn(C)a homogeneous polynomialQ∈HCm[x0,···,xn]such that the common zero set of all coefficients ofQhas codimension at least two.

A moving hypersurfaceQin Pn(C) of degreed(≥1) is defined by

where Id= {(i0,···,in) ∈;i0+···+in=d}, aI∈HCmandωI=We consider the meromorphic mappingQ′:Cm→PN(C),whereN=given by

HereI0<···

The moving hypersurfaces Q = {Q1,···,Qq} is said to be “slow” (with respect tof) if‖TQ′(r) =o(Tf(r)). This is equivalent toLetbe a family of moving hypersurfaces in Pn(C),degQi=di.Assume that

We denote by KQthe smallest subfield of meromorphic function field M which contains C and allwhereaiIt/≡0,i∈{1,···,q},It,Is∈Idi.We say thatfis linearly nondegenerate over KQif there is no nonzero linear formL∈KQ[x0,···,xn] such thatL(f0,···,fn) ≡0,andfis algebraically nondegenerate over KQif there is no nonzero homogeneous polynomialQ∈KQ[x0,···,xn] such thatQ(f0,···,fn)≡0.

LetfandQbe as above. The proximity function offwith respect toQ,denoted bymf(r,Q),is defined by

whereis the Weil function and ‖Q‖ =where Id:= {I= (i0,···,in) ∈;i0+···+in=d}.This definition is independent of the choice of the reduced representation off.

We denote byf∗Qthe pullback of the divisorQbyf.We may see thatf∗Qidentifies with the zero divisorof the functionQ().By Jensen’s fomular, we have

For convenience, we will denoteN(r,f∗Q)=NQ(f)(r).

Theorem 2.1(First Main Theorem)(see [9])Let f:Cm→Pn(C)be a holomorphic map,and let Q be a hypersurface inPn(C)of degree d. If f(Cm) /⊂Q, then for every real number r with0

where O(1)is a constant independent of r.

2.3 Some theorems and lemmas

Letfbe a nonconstant meromorphic map of Cminto Pn(C).Denote by Cfthe set of all non-negative functionsh:CmA→[0,+∞]⊂,which are of the form

wherek,l∈N,g1,···,gl+k∈Kf{0} andA⊂Cm,which may depend ong1,···,gl+k,is an analytic subset of codimension at least two. Then, forh∈Cfwe have

Since Q={Q1,···,Qq} are inN-subgeneral position, we have the following lemma.

Lemma 2.1(see [2, 8, 17])For any Qj1,···,QjN+1∈Q, there exist functions h1,h2∈Cfsuch that

Lemma 2.2(Lemma on Logarithmic Derivative)(see[6])Let f be a nonzero meromorphic function onCm.

Proposition 2.1 (see [5])LetΦ1,···,Φkbe meromorphic functions onCmsuch that{Φ1,···,Φk}is linearly independent overC. Then there exists an admissible set

with≤i−1 (1 ≤i≤k)such that the following are satisfied:

(i) {DαiΦ1,···,DαiΦk}ki=1is linearly independent overM, i.e.,det(DαiΦj)/≡0,

(ii) det(Dαi(hΦj))=hk·det(DαiΦj)for any nonzero meromorphic function h onCm.

Theorem 2.2(see [10, Theorem 2.31])Let f be a linearly nondegenerate meromorphic mapping ofCmintoPn(C)with a reduced representation= (f0,···,fn)and let H1,···,Hqbe q arbitrary hyperplanes inPn(C). Then we have

where α is an admissible set with respect to(as in Proposition2.1)and the maximum is taken over all subsets K⊂{1,···,q}such that{Hj;j∈K}is linearly independent.

3 Proof of Theorem 1.3

Firstly, we may assume thatQ1,···,Qqhave the same degree degQj=dj=d.Set

For eachj, there existsajIj(z), one of the coefficients inQj(z), such thatajIj(z) /≡0.We fix thisajIj, then setand

which is a homogeneous polynomial in KQ[x0,···,xn].By definition of the proximity function and Weil function, we have

forj=1,···,q.

For a fixed pointz∈We may assume that there exists a renumbering {1,···,q} such that

Hence

Let KQbe an arbitrary field over Cmgenerated by a set of meromorphic function on Cm.LetVbe a sub-variety in Pn(C) of dimensionℓdefined by the homogenous idealI(V) ⊂C[x0,···,xn].Denote byIKQ(V) the ideal in KQ[x0,···,xn] generated byI(V).Sincef:Cm→V⊂Pn(C)is algebraically nondegenerate over KQ,there is no homogeneous polynomialP∈KQ[x0,···,xn]/IKQ(V) such thatP(f0,···,fn)≡0.

For a positive integerL,let KQ[x0,···,xn]Lbe the vector space of homogeneous polynomials of degreeL, and letIKQ(V)L:=IKQ(V)∩KQ[x0,···,xn]L. Set

Denote by [g] the projection ofginVL. We have the following basic fact from the theory of Hilbert polynomials (see [15]).

Lemma 3.1M:=dimKQVL=+ρ(L), where ρ(L)is an O(Lℓ−1)function dependingon the variety V. Moreover, there exists an integer L0such that ρ(L)is a polynomial function of L for L>L0.

Next,we prove the following lemma concerning on the hypersurfaces located inN-subgeneral position with indexκ,which plays the role in this paper. The method of it is originally from Quang [9–10].

Lemma 3.2Letbe homogeneous polynomials inKQ[x0,···,xn]of the same degree d≥1, in(weakly)N-subgeneral position with index κ in V. For each point a∈Cmsatisfying the following conditions:

(i)The coefficients ofare holomorphic at a;

such that

ProofWe assume that(1 ≤i≤N+1) has the following form

By the definition of theN-subgeneral position,there exists a pointa∈C such that the following system of equations

has only trivial solution (0,···,0).We may assume thatQi(a)/≡0 for all 1 ≤i≤N+1.

For each homogeneous polynomialQ∈C[x0,···,xn],we denote byDthe fixed hypersurface in Pn(C) defined byQ, i.e.,

where dim ∅=−∞.According to the definition ofN-subgeneral position with indexκ, we have dim

Step 1We will constructas follows. For each irreducible componet Γ of dimensionℓ−κofwe put

By definition,V1Γis a subspace of CN−ℓ+1. Since there existsi∈{κ+ 1,···,N−ℓ+κ+1} such that Γ /⊂Di(a).This implies thatV1Γis a proper subspace of CN−ℓ+1.Since the set of irreducible components of dimensionℓ−κofis at most countable, we have

Hence, there exists(c1(κ+1),···,c1(N−ℓ+κ+1))∈CN−ℓ+1such that Γ/⊂wherefor all irreducible components of dimensionThis clearly implies that

Step 2We will constructas follows. For each irreducible componet Γ′of dimensionℓ−κ−1 ofwe put

ThenV2Γ′is a subspace of CN−ℓ+2. Since dimthere existsi∈{κ+1,···,N−ℓ+κ+2}such that Γ′/⊂Di(a).This implies thatV2Γ′is a proper subspace of CN−ℓ+2.Since the set of irreducible components of dimensionℓ−κ−1 ofis at most countable, we also have

Hence,there exists(c2(κ+1),···,c2(N−ℓ+κ+2))∈CN−ℓ+2such that Γ′/⊂Pκ+2(a),wherefor all irreducible components of dimensionℓ−κ−1 ofThis clearly implies that

Repeat again the above steps, after (ℓ+1 −κ)-th step we get the hypersurfacessatisfying that

wheret=κ+1,···,ℓ+1.

Since there are only finitely many choices ofN+1 polynomials fromthe total number of suchsis finite,so there exists a constantC >0,fort=κ+1,···,ℓand allz∈Cm(excluding all zeros and poles of all(f)), by Lemma 3.2 we can constructsuch that

forκ+1 ≤t≤ℓ,and thus,

Combining the above inequality with (3.1), we have

This gives that ifN−ℓ≤κ,we have

and ifN−ℓ≥κ, we get

whereh′′′∈Cf.Hence, by (3.2)–(3.3), we get

whereh∗=max{h′′,h′′′}∈Cf.

Fix a basis {[φ1],···,[φM]} ofVLwith [φ1],···,[φM]∈KQ[x0,···,xn],and let

For every positive integerLdivided byd,we use the following filtration of the vector spaceVLwith respect toThis is a generalization of Corvaja-Zannier’s filtration [1], see in the three references [3, 12, 17].

We arrange,by the lexicographic order,theℓ-tuples i=(i1,···,iℓ)of non-negative integers and set

Definition 3.1(see [3, 12, 17]) (i)For eachi ∈and non-negative integer L with L≥d‖i‖, denote by IiLthe subspace ofKQ[x0,···,xn]L−d‖i‖consisting of all

such that

for some γe∈KQ[x0,···,xn]L−d‖e‖.

(ii)Denote by Iithe homogeneous ideal inKQ[x0,···,xn]generated by

Remark 3.1(see [3, 12, 17]) From this definition, we have the following properties.

(ii)Ii∩KQ[x0,···,xn]L−d‖i‖=IiL.

(iv)If i1−i2:=(i1,1−i2,1,···,i1,ℓ−i2,ℓ)∈HenceIi2⊂Ii1.

Lemma 3.3(see [3, 12, 17]) {Ii|i ∈}is a finite set.

Denote by

Lemma 3.4(see [3, 12, 17]) (i)There exists a positive integer L0such that, for eachi ∈,ΔiLis independent of L for all L satisfying L−d‖i‖>L0.

(ii)There is an integersuch thatΔiL≤Δfor alli ∈and L satisfies L−d‖i‖>0.

Set Δ0:=

Remark 3.2(see [3, 12, 17]) By (iv) of Remark 3.1, if i −i0∈then Δi≤Δi0.

Now, for an integerLbig enough, divisible byd, we construct the following filtration ofVLwith respect toDenote byτLthe set of i ∈withL−d‖i‖>0,arranged by the lexicographic order. Define the spacesWi=WL,iby

Clearly,W(0,···,0)=KQ[x0,···,xn]LandWi⊃Wi′if i′>i,so{Wi}is a filtration of KQ[x0,···,xn]L.SetW∗i={[g]∈VL|g∈Wi}.Hence,{W∗i} is a filtration ofVL.

Lemma 3.5(see [3, 12, 17])Suppose thati′followsiin the lexicographic order, thenCombining with the notation (3.6), we have

Set

We have the following properties.

Lemma 3.6(see [3, 12, 17]) (i) Δ0=Δifor alli ∈τ0L.

(ii)♯τ0L= 1dℓLℓℓ! +O(Lℓ−1).

(iii) ΔiL=Δdℓfor alli ∈τ0L.

We can choose a basis B ={[ψ1],···,[ψM]} ofVLwith respect to the above filtration. Let[ψs]be an element of the basis,which lies inW∗i/W∗i′,we may writeψs=whereγ∈KQ[x0,···,xn]L−d‖i‖.For every 1 ≤j≤ℓ,we have

(The proof of (3.7) can be found in [12, (3.6)]). Hence, by (3.7) and the definition of the Weil function, we obtain

whereh∗∗∈Cf.The basis {[ψ1],···,[ψM]} can be written as linear formsL1,···,LM(over KQ)in the basis{[φ1],···,[φM]}andSince there are only finitely many choices ofthe collection of all possible linear formsLs(1 ≤s≤M)is a finite set,and denote it by L:=(Λ<+∞).It is easy to see that KL⊂KQ.

Lemma 3.7(Product to the Sum Estimate, see [10])Let H1,···,Hqbe hyperplanes inPn(C)located in general position. Denote by T the set of all injective maps µ:{0,1,···,n}→{1,···,q}. Then

holds for all r outside a set with finite Lebesgue measure.

By (3.4), (3.8) and Lemma 3.7, taking integration on the sphere of radiusr, we have

for allroutside a set with finite Lebesgue measure, where the set K ranges over all subsets of{1,···,Λ} such that the linear forms {Lj}j∈Kare linearly independent.

By Theorem 2.2, we have, for anyε>0,

holds for allroutside a setEwith finite Lebesgue measure. Takingε=12in (3.10), and using(3.5) and (3.9), we obtain

holds for allr/∈E,whereWis the Wronskian ofF1,···,FM.

TakeLlarge enough such thatε 0 is any given constant in the theorem. Then we have

holds for allr/∈E.

And by the first main theorem, (3.11) can be written

Secondly, for the general case whenever allQjmay not have the same degree, consideringinstead ofQj,we haveNf(r,Qj)=Then the theorem is proved immediately.

4 Proof of Theorem 1.4

ReplacingQibyif necessary with the note that

we may assume that all hypersurfacesQi(1 ≤i≤q) are of the same degreed.We may also assume thatq >

Consider a reduced representation=(f0,···,fn):C →Cn+1off.We also note that

Then without loss of generality we may assume thatQi∈Kf[x0,···,xn].

We set

For eachI= (i1,···,iN+1) ∈I,we denote byPI1,···,PI(n+1)the moving hypersurfaces obtained in Lemma 3.2 with respect to the family of moving hypersurfaces {Qi1,···,QiN+1}.It is easy to see that there exists a positive functionh∈Cfsuch that

for allI∈I andω=(ω0,···,ωn)∈Cn+1.

For a fixed pointwe may assume that

LetI=(i1,···,iN+1).SincePI1,···,PI(n+1)are in weakly general position, there exist functionsg0,g∈Cf,which may be chosen independent ofIandz,such that

Therefore, we have

whereh1=gq−N(z)hn−κ(z), I= (i1,···,iN+1).Choose a functionζin Cfwhich is common for allI∈I,such that

We first consider ifN−n≤κ,and have by the inequality above that,

whereh2=ζ(z)2n−Nh1∈Cf,however ifN−n≥κ, we get

whereh3=

Thus by (4.2)–(4.3), we get

whereh∗=max{h2,h3}∈Cf.

Hence, by taking logarithms in the both sides of (4.4), we can obtain

Now,for each non-negative integerL,we denote byVLthe vector space(over KQ)consisting of all homogeneous polynomials of degreeLin KQ[x0,···,xn]and the zero polynomial. Denote by (PI1,···,PIn) the ideal in KQ[x0,···,xn] generated byPI1,···,PIn.

Lemma 4.1(see [2, Proposition 3.3])Let{Pi}qi=1(q≥n+1)be a set of homogeneouspolynomials of common degree d≥1inKf[x0,···,xn]in weakly general position. Then for any nonnegative integer L and for any J:= {j1,···,jn} ⊂{1,···,q}, the dimension of thevector spaceis equal to the number of n-tuples(s1,···,sn) ∈Nn0such that s1+···+sn≤L and0 ≤s1,···,sn≤d−1. In particular, for all L≥n(d−1), we have

Now, for each positive integerLbig enough, divided byd,and i = (i1,···,in) ∈Nn0withwe set

It is clear thatWL,(0,···,0)=VLandWL,i⊃WL,i′if i

Lemma 4.2Leti = (i1,···,in),i′= (i′1,···,i′n) ∈Nn0. Suppose thati′followsiin the lexicographic ordering and dσ(i)

This lemma yields that

Fix a numberLlarge enough (chosen later). Setu=uL:= dimVL=We assume that

whereWL,is+1followsWL,isin the ordering and iK=,0,···,0.It is easy to see thatKis the number ofn-tuples (i1,···,in) withij≥0 andi1+···+in≤.Then we have

For eachk∈{1,···,K−1},we setmIk=dimandmIK= 1.Then by Lemma 4.1,mIkdoes not depend on {PI1,···,PIn} andk,but only onσ(ik).Hence, we setmk:=mIk.We also note that by Lemma 4.1,

for allkwithL−dσ(ik)≥n(d−1)(it is equivalent toσ(ik)≤Ld−n).

From the above filtration, we may choose a basis {ψI1,···,ψIu} ofVLsuch that is a basis ofWL,is.For eachk∈{1,···,K} andl∈{u−(mk+ ··· +mK) + 1,···,u−(mk+1+···+mK)},we may write

We may choosehlto be a monomial.

We have the following estimates: Firstly, we see that

Note that, by the symmetry (i1,···,in) →(iσ(1),···,iσ(n)) withσ∈S(n),does not depend ons.We set

Then we have

wherecl∈Cf,which does not depend onfandz.Taking the product on both sides of the above inequalities over allland then taking logarithms, we obtain

wherecI=By (4.7), it gives

i.e.,

Setc0=

Combining (4.8) with (4.5), we obtain that

We now write

where ILis the set of all (n+1)-tuplesJ= (i0,···,in) with=L,xJ=xj0

0 ···xjnnandl∈{1,···,u}.For eachl,we fix an indexJIl∈Jsuch that/≡0.Define

Set Φ= {µIlJ;I⊂{1,···,q},♯I=n,1 ≤l≤u,J∈IL}.Note that 1 ∈Φ.LetB=♯Φ.We see thatB≤=For each positive integerl,we denote by L(Φ(l)) the linear span over C of the set

It is easy to see that

We may choose a positive integerpsuch that

and

Therefore, we have

This is a contradiction.

We fix a positive integerpsatisfying the above condition. Puts= dim L(Φ(p)) andt=dim L(Φ(p+1)).Let {b1,···,bt} be an C-basis of L(Φ(p+ 1)) such that {b1,···,bs} be a C-basis of L(Φ(p)).

For eachl∈1,···,u,we set

For eachJ∈IL,we consider homogeneous polynomialsφJ(x0,···,xn)=xJ.LetFbe a meromorphic mapping of Cminto Ptu−1(C)with a reduced representationwherehis a nonzero meromorphic function on Cm.We see that

Sincefis assumed to be algebraically nondegenerate over KQ,Fis linearly nondegenerate over C.We see that there exist nonzero functionsc1,c2∈Cfsuch that

For eachl∈1,···,u,1 ≤i≤s,we consider the linear formLIilinxJsuch that

Sincefis algebraically nondegenerate over KQ,it is easy to see thatl≤M} is linearly independent over C,and so is {LIil();1 ≤i≤s,1 ≤l≤u}.This yields that {LIil;1 ≤i≤s,1 ≤l≤u} is linearly independent over C.

For every pointzwhich is neither zero nor pole of anyhbiψIl(f),we also see that

wherec3,c4are nonzero functions in Cf,not depending onfandI, but onCombining this inequality and (4.9), we obtain that

for allzoutside an analytic subset of Cm.

By Theorem 2.2, we have

Integrating both sides of (4.10) and using (4.11), we obtain that

We can estimate the following quantity using the method of Quang [8],

thus we can get

From this inequality and (4.12) with a note thatTF(r)=LTf(r)+o(Tf(r)),we have

Now we give some estimates forA,tands.For eachIk=(i1k,···,ink)withσ(ik)≤Ld−n,we set

Since the number of nonnegative integerp-tuples with summation ≤Iis the same as the number of nonnegative integer (p+1)-tuples with summation exactly equal toI∈Z,which isand since the sum below is independent ofs,we have

Now, for every positive numberwe have

We choseThenLis divisible bydand we have

Therefore, using (4.14)–(4.15), we have

Then we have

Combining (4.13) and (4.16), we get

Here we note that

By these estimates and by (4.17), we obtain

The theorem is proved.

AcknowledgementThe authors are very grateful to the anonymous referee for some valuable suggestions and comments.


登录APP查看全文