This paper proves the D'Angelo conjecture for a specific class of rational holomorphic maps between unit balls, establishing degree bounds in the third gap interval for dimensions n ≥ 7.
Contribution
It demonstrates that the degree of rational proper holomorphic maps from to in the third gap interval is at most 3, confirming the conjecture in this case.
Findings
01
D'Angelo conjecture holds in the third gap interval.
02
Degree of maps from to is at most 3 for n 7.
03
Validates the conjecture for a broad class of rational maps.
Abstract
We show the D'Angelo conjecture holds in the third gap interval. More precisely, we prove that the degree of any rational proper holomorphic map from Bn to B4n−6 with n≥7 is not more than 3.
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsHolomorphic and Operator Theory · Meromorphic and Entire Functions · Geometry and complex manifolds
Full text
D’Angelo conjecture in the third gap interval
Shanyu Ji111Supported in part by NSFC-11571260
and Wanke Yin222Supported in part by NSFC-11722110 and
NSFC-11571260.
Abstract
We show the D’Angelo conjecture holds in the third gap
interval. More precisely, we prove that the degree of any rational
proper holomorphic map from Bn to B4n−6
with n≥7 is not more than 3.
1 Introduction
Let Bn={z∈Cn∣∣z∣<1} be the unit ball in Cn and
denote by Rat(Bn,BN) the set of all proper holomorphic
rational maps from Bn to BN. We say that f,g∈Rat(Bn,BN) are holomorphically equivalent (or equivalent, for short) if there are σ∈Aut(Bn) and
τ∈Aut(BN) such that f=τ∘g∘σ. By a
well-known result of Cima-Suffridge [CS], F extends
holomorphically across the boundary ∂Bn.
For the equal dimensional case N=n, Alexander [A77] proved
that Rat(Bn,Bn) must be automorphisms for n>1.
Subsequently, much effort has been paid to the classification of
Rat(Bn,BN) with N>n. When N/n is not too large, it
turns out that the maps are equivalent to relatively simple ones. In
fact, the classification problem had been done for N≤3n−3 and
the maps turn out to be all monomial maps. The systematic
investigations on the precise classification of Rat(Bn,BN)
can be found in the work of
[F82, Hu99, HJ01, Ha05, HJX06, CJY18, JY18], etc. In [FHJZ10],
Faran-Huang-Ji-Zhang constructed a family of maps in Rat(Bn,B3n−2), which cannot be equivalent to any polynomial maps.
This indicates that the maps could be quite
complicated when N≥3n−2.
To study maps in Rat(Bn,BN) , there are two geometric
problems which are of fundamental importance. The first one is the
D’Angelo conjecture.
For any rational holomorphic map H=Q(P1,...,Pm) on
Cn where Pj,Q are holomorphic polynomials with (P1,...,Pm,Q)=1, the degree of H is defined, as in algebraic
geometry, to be deg(H):=max{deg(Pj),deg(Q),1≤j≤m}.
The D’Angelo conjecture is as follows: For any F∈Rat(Bn,BN), it should have
[TABLE]
There are several partial results supporting this conjecture. The
conjecture is true for all monomial maps, as demonstrated by
D’Angelo-Kos-Riehl [DKR03] for the case n=2 and
by Lebl-Peter [LP12] for the case n≥3. If F is a
rational map with geometric rank one, this conjecture was proved in
[HJX06, corollary 1.3]. If the conjecture is proved, it would
be sharp due to the known examples. Also, it is proved that
deg(F)≤2(2n−3)N(N−1) holds for any F∈Rat(Bn,BN) with n=2 in [Me06] and with n≥2 in [DL09].
Another geometric problem is the gap conjecture.
For any integer k with 1≤k≤K(n) where K(n):=max{t∈Z+∣2t(t+1)<n}, we recall the gap interval
(cf. [HJY09])
[TABLE]
The gap conjecture, first raised in [HJY09], is stated
as follows: Any proper holomorphic rational map F∈Rat(Bn,BN) where n≥3 is equivalent to a map of the form (G,0′)
where G∈Rat(Bn,BN′) where N′<N if and only if N∈Ik for some 1≤k≤K(n). The gap conjecture for the
cases of k=1,2,3 have been proved [Hu99, HJ01, HJY14].
Recently, P. Ebenfelt [Eb16] proposed a SOS conjecture (i.e.,
the Sums of Squares of Polynomial conjecture) and proved that if
the SOS conjecture is true, then it implies the gap conjecture.
The first gap interval is I1=(n,2n−1). Huang [Hu99]
showed that deg(F)=1 if N∈I1. When N=2n−1, it was
proved by Faran [F82] that deg(F)≤3 for n=2 and by
Huang-Ji [HJ01] that deg(F)≤2 for n>2. The second gap
interval is I2=(2n,3n−3). When N≤3n−3 (and n≥4), we know from [AHJY15] that deg(F)≤2. These results
confirm the D’Angelo conjecture for the first and the second gap
intervals.
The third gap interval is I3=(3n,4n−6). If D’Angelo
conjecture is true, we would have deg(F)≤3 for any F∈Rat(Bn,B4n−6) because deg(F)≤n−14n−6−1=4−n−13. This is confirmed by our main
result of this paper as follows.
Theorem 1.1**.**
If F∈Rat(Bn,B4n−6) with n≥7, then deg(F)≤3.
The rest of the paper is organized as follows. In Section 2, we
introduced some known properties for Rat(Hn,HN), especially
for maps of geometric rank 2. Section 3 was devoted to the proof
of our main theorem assuming Proposition 3.2. In Sections
4-7, we gave a detailed proof of Proposition 3.2 according
to four different cases.
2 Preliminaries
Let Hn={(z,w)∈Cn−1×C∣Im(w)>∣z∣2}
be the Siegel upper half space and denote by Rat(Hn,HN) the
set of all proper holomorphic rational maps from Hn to HN.
By the Cayley transform, we can identify Bn with Hn and
identify Rat(Bn,BN) with Rat(Hn,HN). In what
follows, we will prove Theorem 1.1 through the properties
of Rat(Hn,HN).
Let F=(f,ϕ,g)=(f,g)=(f1,⋯, fn−1,
ϕ1,⋯, ϕN−n,g)∈Rat(Hn,HN). For each
p∈∂Hn, define σp0∈Aut(Hn) and
τpF∈Aut(HN) as follows:
[TABLE]
Then
F is equivalent to Fp:=τpF∘F∘σp0=(fp,ϕp,gp) and Fp(0)=0. In [Hu99], Huang
constructed an automorphism τp∗∗∈Aut0(HN) such that
Fp∗∗:=τp∗∗∘Fp satisfies the following
normalization:
[TABLE]
Write
A(p):=−2i(∂zj∂w∂2(fp)l∗∗∣0)1≤j,l≤n−1. The geometric rank of F at
p is defined to be the rank of the (n−1)×(n−1) matrix
A(p), which is denoted by RkF(p). Now we define the
geometric rank of F to be κ0(F)=maxp∈∂HnRkF(p).
When a map in Rat(Hn,HN) is not of full rank (i.e.,
κ0≤n−2), by the works of [Hu03] and [HJX06],
it can further be normalized to the following form:
Theorem 2.1**.**
Suppose that F∈Rat(Hn,HN)
has geometric rank 1≤κ0≤n−2 with F(0)=0. Then there are
σ∈Aut(Hn) and
τ∈Aut(HN) such that
τ∘F∘σ takes
the following form, which is still denoted by F=(f,ϕ,g) for
convenience of notation:
[TABLE]
Here, for 1≤κ0≤n−2, we write S=S0∪S1, the index set for all components of ϕ, where
S0={(j,l):1≤j≤κ0,1≤l≤n−1,j≤l}, {\cal S}_{1}=\Big{\{}(j,l):j=\kappa_{0}+1,\kappa_{0}+1\leq l\leq\kappa_{0}+N-n-\frac{(2n-\kappa_{0}-1)\kappa_{0}}{2}\Big{\}}, and
[TABLE]
Write ϕ=(Φ0,Φ1) where Φ0=(ϕlk)(l,k)∈S0, Φ1=(ϕlj)(l,k)∈S1. Define
Φ0(1,1)(z)=∑j=1κ0ejzj,
Φ1(1,1)(z)=∑j=1κ0ek^zj, ej∗=(ej,e^j), and ξj(z)=ej⋅Φ0(2,0)(z). Here we
use notation Φ(s,t)(z) to denote a homogeneous polynomial of
degree s in z, and
[TABLE]
Let F be as in Theorem 2.1. By [HJY14, Corollary
3.4], we have the following explicit expression for
Φ1(3,0)(z).
Lemma 2.2**.**
Let κ0≥2 and (κ0+1)n−κ0≤N≤(κ0+2)n−κ02−2. Then, after applying
a unitary transformation to the Φ1-components,
F is equivalent to another map such that the new map
(still denoted as F) has the property
[TABLE]
We call a map satisfying Theorem 2.1 and
(2.3) a normalized map, and denote it by
F∗∗∗.
From the above lemma and [HJY14, (3.7)], we obtain
[TABLE]
where ξ=(ξ1,...,ξκ0)=(e1,...,eκ0)⋅Φ0(2,0).
Next, we show more properties for F∈Rat(Hn,H4n−6)
with n≥7 and geometric rank κ0=2.
By [HJY14, (4.3)], we have 2iμ1an(1)(ϵ)+f1(1,2)(ϵ,0,...,0)=0,2iμ2bn(1)(ϵ)+f2(1,2)(ϵ,0,...,0)=0, and ϕ(1,2)(ϵ,0,...,0)+e1∗an(1)(ϵ)+e2∗bn(1)(ϵ)=0. Thus
[TABLE]
Next we recall the definitions and the properties of the degeneracy
ranks. These invariants are introduced by Lamel [La01] and
Ebenfelt-Huang-Zaitsev [EHZ04].
Let F∈Rat(Hn,HN) and p∈∂Hn. Write Lj=∂zj∂+2izj∂w∂. Then {Lj}1≤j≤n−1 forms a basis of tangent vector fields of (1,0) on
∂Hn. Denote by ρ^(Z,Z) the defining function of
the real hypersurface ∂HN, and denote by ρ^Z:=∂ρ^ the complex gradient of ρ. Now we define
an increasing sequence of linear subspaces Ek(p)⊂CN as follows:
[TABLE]
Define d1(p):=0 and dk(p):=dimCEk(p)/E1(p). By moving p to a nearby point p0 if
necessary, we may assume that all dl(p) are locally constant near
p0 and
[TABLE]
for some l0 with 1≤l0≤N−n+1. By [EHZ04], l0 is
called the degeneracy rank of F, and dl0 is called
the degeneracy dimension of F. These definitions depend on
the open subset U. degeneracy rankl0 of F is defined
to be the
minimization of l0 among all these open sets. Then we have the
following degree estimates when both κ0 and l0 are
small.
Theorem 2.3**.**
([CJY18, Theorem 1.4]) Let F∈Rat(Bn,BN) with 5≤n≤N. Suppose that κ0≤2 and l0≤2. Then deg(F)≤2.
Finally, we finish this section by recalling a lemma of [HJ01],
which gives a reduction for the degree estimates of a rational
holomorphic map.
For any point q=(q,qn)∈Cn, the Segre family of
∂Hn is a family given by Qq:={(ζ,τ)∈Cn−1×C∣τ=qn+2i⟨ζ,q⟩)}. Let F:Cn→CN be a rational holomorphic map.
Then the restriction F∣Qq(ζ)=F(ζ,qn+2i⟨ζ,q⟩) is a rational holomorphic map in ζ∈Cn−1.
Lemma 2.4**.**
([HJ01, lemmas 5.3-5.4]) Let F=Q(P1,...,PN) be a rational holomorphic map from Cn into CN
where Pj and Q are as above. Suppose that there exists a fixed
positive integer k such that one of the following two conditions
is satisfied:
This section is devoted to the proof of Theorem 1.1,
assuming Proposition 3.2, whose proof will be given in the
end of this section and in §4−7.
We start with some notations. For any F∈Rat(Hn,H4n−6)
and q∈∂Hn, the map associates to a normalized map
Fq∗∗∗. We write Fq∗∗∗=(fq∗∗∗,ϕq∗∗∗,gq∗∗∗) and \phi_{q}^{***}=\big{(}(\Phi_{q}^{***})_{0},(\Phi_{q}^{***})_{1}\big{)},
which are decomposed similar to those of F in Theorem
2.1.
Lemma 3.1**.**
Let F∈Rat(Hn,H4n−6) be a
normalized map with κ0=2 and n≥7. Then
[TABLE]
for any index α, where Dα=∂z1α1⋯∂znαn∂∣α∣ is the standard
differential operator. If, in addition,
\big{(}\Phi^{***}_{q}\big{)}_{1}^{(3,0)}(z)\equiv 0 for any
q in
an open neighborhood of [math] in ∂Hn, then for any index
α,
[TABLE]
Proof.
By [HJY14, Theorem 4.1 and (5.3)], we have
span∣β∣≤4{LβF∣0}≤span∣α∣≤3{LαF∣0}. By [HJY14, claim (5.5)], for any
p∈∂Hn, the associated map Fp satisfies
[TABLE]
Since Lα(Fp)∣0=LαF(p), it implies
span∣β∣≤4{LβF∣p}≤span∣α∣≤3{LαF∣p}. Then by similar argument as that in
[HJY14, p. 139], through applying Lj,Lk
to F, we obtain (3.1). If, in addition,
(Φq∗∗∗)1(3,0)(z)≡0 for any
q in
an open neighborhood of [math] in ∂Hn, by the last paragraph of
[HJY14, p. 139], we get (3.2).
∎
As mentioned in §2, by the Cayley transform, we can
identify Rat(Bn,BN) with Rat(Hn,HN). It suffices for us to show for
any F∈Rat(Hn,H4n−6) with n≥7,
we have deg(F)≤3.
By [Hu03, Theorem 1.1], we have N≥n+2(2n−κ0−1)κ0. Hence for N=4n−6, its
geometric rank κ0≤3. By [Hu99], we can suppose 1≤κ0≤3, otherwise the map must be linear fractional.
If F∈Rat(Hn,H4n−6) with n≥3 and κ0=1, by [HJX06, Corollary
1.3], we have deg(F)≤3. If F∈Rat(Hn,H4n−6) with n≥5 and κ0=3, by [JY18, Theorem
1.1], we have deg(F)≤2. Therefore, we assume κ0=2
for F in the rest proof of Theorem 1.1.
Let F∈Rat(Hn,H4n−6) with n≥7 and κ0=2.
For any q∈∂Hn, we study the associated map (Fq)∗∗∗.
There are two cases to consider.
Case I:Fq∗∗∗ satisfies
\big{(}\Phi_{q}^{***}\big{)}_{1}^{(3,0)}(z)\equiv 0 for any q in a
neighborhood of [math] in ∂Hn. Recall that Fq∗∗∗ is a
normalized map with geometric rank 2. Hence by (3.2), it implies that the degeneracy rank l0≤2. Then we apply
Theorem 2.3 to conclude deg(F)≤2.
Case 2:Fq0∗∗∗ satisfies that
\big{(}\Phi_{q_{0}}^{***}\big{)}_{1}^{(3,0)}(z)\not\equiv 0 for some
q0∈∂Hn. As in Case I, Fq0∗∗∗ is a normalized map
with geometric rank 2. Now we need the following proposition.
Proposition 3.2**.**
Let F∈Rat(Hn,H4n−6)(n≥7) be a normalized map, whose geometric rank
κ0=2. Suppose that (Φ1)(3,0)(z)≡0.Then
[TABLE]
The proof of Proposition 3.2 will be postponed to the next
section. Admitting this proposition temporarily, we continue the
proof of Theorem 1.1.
Notice that (Φq∗∗∗)1(3,0)(z) is smooth with respect
to q. From (Φq0∗∗∗)1(3,0)(z)≡0, we know
(Φq∗∗∗)1(3,0)(z)≡0 for q in a small
neighborhood of q0. Applying Proposition 3.2 with F
replaced by Fq∗∗∗ for any q∈∂Hn in a neighborhood of
q0, we conclude that deg(Fq∗∗∗(z,0))≤3. Since maps in
both Aut(∂Hn) and Aut(∂HN) are linear fractional, we
know deg(Fq(z,0))≤3. Now we use Lemma 2.4 to
complete the proof of Theorem 1.1.
∎
By [HJX06, Lemma 2.3], we know fl(z,0)=zl. From Lemma
2.2, Φ1=(ϕj,k)(j,k)∈S1=(ϕ33,ϕ34,...,ϕ3(n−2)) satisfies Φ1(3)(z,0)=(ϕ33(3)(z,0),0,...,0). Then we apply Lemma
3.1 to know
[TABLE]
Hence it suffices to prove
[TABLE]
By our notation,
ϕ33(3,0)=(ϕ33(j1I1+jn−1In−1))j1+⋯+jn−1=3,j1+j2≥1. From (2.5) and (3.4),
(Φ1)33(3,0)(z)≡0 if and only if
[TABLE]
Thus it is suffices for us to consider the following cases:
[TABLE]
In the rest of the paper, we will obtain an explicit expression for
ϕ33(z,0) and ϕjk(z,0) with (j,k)∈S0,
from which we obtain (3.5).
We start with the following Chern-Moser equation.
[TABLE]
By complexification, we write
[TABLE]
Applying Lj:=∂zj∂+2iχj∂w∂ for z=0 and w=η=0 to the both sides of the above
identity, we obtain
[TABLE]
[TABLE]
We notice that the index set S in (3.8) is
replaced by S0 in (3.9) and (3.10)
because ϕst(2)(0,0)=0 for any (s,t)∈S1. The
proof of Proposition 3.2 will be completed in Sections
4−7 according to the different cases in (LABEL:ABCD).
∎
Similarly by considering B2,23ϕ(χ,0)t=A2,23,
we obtain
[TABLE]
where
[TABLE]
Substituting (LABEL:rel) into B2,33ϕ(χ,0)t=A2,33, we yield
[TABLE]
We notice that the polynomial C2=0 because e2,11=0.
Here we used the fact that ϕ33(3I1)=μ2(μ1+μ2)2μ1e2,11=0 in Case A.
Then divided by C2, we obtain from above
[TABLE]
and hence
[TABLE]
Then (4.16) and (4.17) further take the form at (χ,0):
[TABLE]
It implies that ϕ13(χ,0) and ϕ23(χ,0) take the form χ3P2(2)(χ)P1(2)(χ)
where P1(2)(χ) and P2(2)(χ) are polynomials of
degree 2. Substituting these forms back to (LABEL:rel) and (4.19), we conclude that (3.5) is proved and hence
deg(F(z,0))≤3.
This completes the proof of Proposition 3.2 for Case A.
Case A*′* can be similarly proved.
∎
A direct computation shows that L12L2=∂z12∂z2∂3+2iχ2∂z12∂w∂3+4iχ1∂z1∂z2∂w∂3+4iχ1⋅2iχ2∂z1∂w2∂3+(2iχ1)2∂z2∂w2∂3+(2iχ1)2⋅2iχ2∂w3∂3. Thus
and A1=(A1,11,A1,12,A1,22,A1,1α,A1,2α,A1,33)t where A1,33=−L12L2f1∣(0,0)χ1−L12L2f2∣(0,0)χ2=−(D1+D2)χ1−(D3+D4)χ2, and
B1,33=(L12L2ϕhl∣(0,0),L12L2ϕhα∣(0,0),L12L2ϕ33∣(0,0)). Recall
L12L2ϕ∣(0,0)=2ϕ(2I1+I2)+2iχ2⋅2ϕ(2I1+In)+4iχ1ϕ(I1+I2+In)+4iχ1⋅2iχ2⋅2ϕ(I1+2In)+(2iχ1)2⋅2ϕ(I2+2In). We calculate B1,33 in details as
follows.
[TABLE]
[TABLE]
[TABLE]
[TABLE]
[TABLE]
for j=1,2, and
[TABLE]
[TABLE]
We write B2=(Id−G1)B1 and A2=(Id−G1)A1 so that
[TABLE]
where
[TABLE]
By the construction of B2 and A2, we see that
(LABEL:rel)(4.16) and (4.17) still hold. We further
calculate:
[TABLE]
and
[TABLE]
We turn to B2,33ϕ(χ,0)t=A2,33 to
have
[TABLE]
Substituting (LABEL:rel) to this equation, we obtain
[TABLE]
A quick simplification gives
[TABLE]
Since ϕ33(2I1+I2)=0 in Case B, the polynomial
μ2χ1D2−μ1χ2D3=0 because of (2.5). Thus
[TABLE]
Hence we obtain the same formula (4.19). As in case A, we
show deg(F)≤3. The proof for the case B*′* is similar to the
case B.
∎
Recall L1L2Lj=∂z1z2∂zj∂3+2iχ1∂z2zj∂w∂3+2iχ2∂z1∂zj∂w∂3+2iχj∂z1∂z2∂w∂3−4χ2χj∂z1∂w2∂3−4χ1χj∂z2∂w2∂3−4χ1χ2∂zj∂w2∂3−8iχ1χ2χj∂w3∂3. Then
and A1=(A1,11,A1,12,A1,22,A1,1α,A1,2α,A1,33)t where
A1,33=−L1L2Ljf1∣(0,0)χ1−L1L2Ljf2∣(0,0)χ2=−(D1+D2+D3)χ1−(D4+D5+D6)χ2, and
B1,33=(L1L2Ljϕhl∣(0,0),L1L2Ljϕhα∣(0,0),L1L2Ljϕ33∣(0,0)). Recall L1L2Ljϕ∣(0,0)=ϕ(I1+I2+Ij)+2iχ1ϕ(I2+Ij+In)+2iχ2ϕ(I1+Ij+In)+2iχjϕ(I1+I2+In)−8χ2χjϕ(I1+2In)−8χ1χjϕ(I2+2In). Then
[TABLE]
[TABLE]
[TABLE]
[TABLE]
[TABLE]
[TABLE]
[TABLE]
[TABLE]
[TABLE]
We write B2=(Id−G1)B1 and
A2=(Id−G1)A1 so that
[TABLE]
where
[TABLE]
By the construction of B2 and A2, we see
that (LABEL:rel)(4.16) and (4.17) still hold. We further
calculate:
[TABLE]
and
[TABLE]
We turn to B2,33ϕ(χ,0)t=A2,33 to have
[TABLE]
Substituting (LABEL:rel) to this equation, we obtain
[TABLE]
A quick simplification gives
[TABLE]
Since ϕ33(I1+I2+Ij)=0 in Case C, the polynomial
μ2χ1D2−μ1χ2D4=0 because of
(2.5). Thus
[TABLE]
Hence we obtain the same formula (4.19). As in Case A, we can
further get deg(F)≤3.
∎
Here we used the fact that f1(Ij+2In)=f2(Ij+2In)=0 due
to Theorem 2.1. A direct computation shows that
L12Lj=∂z12∂zj∂3+2iχj∂z12∂w∂3+4iχ1∂z1∂zj∂w∂3−8χ1χj∂z1∂w2∂3−4χ12∂zj∂w2∂3−8iχ12χj∂w3∂3. Thus
and A1=(A1,11,A1,12,A1,22,A1,1α,A1,2α,A1,33)t where
A1,33=−L12Ljf1∣(0,0)χ1−L12Ljf2∣(0,0)χ2=−(D1+D2)χ1−(D3+D4)χ2, and B1,33=(L12Ljϕhl∣(0,0),L12Ljϕhα∣(0,0),L12Ljϕ33∣(0,0)).
We also have L12Ljϕ∣(0,0)=2ϕ(2I1+Ij)+2iχj⋅2ϕ(2I1+In)+4iχ1ϕ(I1+Ij+In)−8χ1χj⋅2ϕ(I1+2In)−4χ12⋅2ϕ(Ij+2In) so that
[TABLE]
[TABLE]
[TABLE]
[TABLE]
[TABLE]
[TABLE]
[TABLE]
We write B2:=(Id−G1)B1 and
A2:=(Id−G1)A1 so that
[TABLE]
where
[TABLE]
By the construction of B2 and A2, we see
that (LABEL:rel)(4.16) and (4.17) still hold. We further
calculate:
[TABLE]
and
[TABLE]
We turn to B2,33ϕ(χ,0)t=A2,33 to have
[TABLE]
Substituting (LABEL:rel) to this equation, we obtain at (χ,0)
[TABLE]
A quick simplification gives at (χ,0)
[TABLE]
Since ϕ33(2I1+Ij)=0 in Case D, the polynomial
μ1D3=0 because of (2.5). Hence we
obtain the same formula (4.19). As in the case A, we can
further get deg(F)≤3. The proof for the case D*′* is similar to
the case D.
∎
Therefore the proofs of Proposition 3.2 and Theorem
1.1 are complete.
Bibliography26
The reference list from the paper itself. Each links out to its DOI / PubMed record.
1[AHJY 15] J. Andrews, X. Huang, S. Ji and W. Yin, Mapping 𝔹 n superscript 𝔹 𝑛 {\mathbb{B}}^{n} into 𝔹 3 n − 3 superscript 𝔹 3 𝑛 3 {\mathbb{B}}^{3n-3} , Comm. Anal. Geom. 24 (2016), no. 2, 279–300.
2[A 77] H. Alexander, Proper holomorphic maps in 𝐂 n superscript 𝐂 𝑛 {\mathbf{C}}^{n} , Indiana Univ. Math. Journal 26 (1977), 137–146.
3[CJY 18] X. Cheng, S. Ji and W. Yin, Mappings between balls with geometric and degeneracy rank two, Sci. China Math. (2018). https://doi.org/10.1007/s 11425-017-9214-6
4[CS] J. Cima and T. J. Suffridge, Boundary behavior of rational proper maps, Duke Math. J. 60 (1990), 135–138.
5[DA 93] J. P. D’Angelo, Several complex variables and the geometry of real hypersurfaces, Studies in Advanced Mathematics, CRC press, 1993.
6[DKR 03] John P. D’Angelo, Simon Kos, and Emily Riehl, A sharp bound for the degree of proper monomial mappings between balls, J. Geom. Anal. 13 (2003), no. 4, 581–593.
7[DL 09] J. P. D’Angelo and J. Lebl, On the complexity of proper mappings between balls, Complex Variables and Elliptic Equations 54 (2009), no. 3, 187–204.
8[Eb 13] P. Ebenfelt, Partial rigidity of degenerate CR embeddings into spheres, Adv. Math. 239 (2013), 72–96.
