We obtain weighted integral representations for spaces of functions holomorphic in the unit ball Bn and belonging to area-integrable weighted Lp-classes with “anisotropic” weight function of the type ∏i=1n(1−|w1|2−|w2|2−⋯−|wi|2)αi, w=(w1,w2,…,wn)∈Bn. The corresponding kernels of these representations are estimated, written in an integral form, and even written out in an explicit form (for n=2).
1. Introduction
Denote by Bn the unit ball in the complex n-dimensional space Cn:Bn={w∈Cn:|w|<1}. For 1≤p<+∞ and α>-1, denote by Hαp(Bn) the space of all functions f holomorphic in Bn and satisfying the condition
∫Bn|f(w)|p(1-|w|2)αdm(w)<+∞,
where dm is the Lebesgue measure in Cn≡R2n. Further, for a complex number β with Reβ>-1, put
cn(β)=Γ(n+1+β)πn⋅Γ(1+β).
We have the following theorem.
Theorem 1.1.
Assume that 1≤p<+∞,α>-1, and that the complex number β satisfies the condition
Reβ≥α,p=1,Reβ>α+1p-1,1<p<∞.
Then each function f∈Hαp(Bn) admits the following integral representations:
f(z)=cn(β)⋅∫Bnf(w)(1-|w|2)β(1-〈z,w〉)n+1+βdm(w),z∈Bn,f(0)¯=cn(β)⋅∫Bnf(w)¯(1-|w|2)β(1-〈z,w〉)n+1+βdm(w),z∈Bn,
where 〈·,·〉 is the Hermitean inner product in Cn.
For n=1, that is, for the case of the unit disc D⊂C, this theorem was established in [1, 2], where the formulas (1.4) are important in the theory of factorization of meromorphic functions in the unit disc.
For n>1, the theorem was proved in [3] (when α=0) and in [4, 5] (when α>-1).
In monographs [6, 7], one can find numerous applications of the formulas (1.4) in the complex analysis.
In the present paper, we generalize Theorem 1.1 in the following way.
Assume that 1≤p<+∞ and α=(α1,…,αn)∈Rn satisfies the conditions
αn>-1,αn+αn-1>-2,αn+αn-1+αn-2>-3,⋮αn+αn-1+αn-2+⋯+α2>-(n-1),αn+αn-1+αn-2+⋯+α2+α1>-n.
Then we introduce the spaces Hαp(Bn) of functions f holomorphic in Bn and satisfying the condition
∫Bn|f(w)|p⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)<+∞.
Section 3 contains detailed investigation of these spaces with “anisotropic” weight function.
For these “anisotropic” spaces, similarities of the integral representations (1.4) are obtained, but this time a special kernels Sβ1β2⋯βn(z;w)≡Sβ(z;w) (where βi,1≤i≤n, are associated with αi,1≤i≤n, and p in a special way) appear instead of cn(β)·(1-〈z,w〉)-(n+1+β) (Theorem 4.7). Theorem 4.5 gives the description (in a multiple series form) and the main properties of these kernels. Theorem 4.8 makes it possible to represent the kernels Sβ as integrals taken over [0;1]n-1⊂Rn. Finally, in the special case n=2 we write out these kernels in an explicit form (see Theorem 4.12).
2. Preliminaries
In this section, we present several well-known facts which will be used in what follows.
Fact 1.
For α>-1 and k=0,1,2,…, put
J(1)(α;k)=∫D|w|2k(1-|w|2)αdm(w),
then
J(1)(α;k)=πΓ(k+1)Γ(α+1)Γ(k+α+2).
Moreover, if α>-1 and k,l=0,1,2,…(k≠l), then
∫Dwkw¯l(1-|w|2)αdm(w)=0.
Remark 2.1.
Assume that ρ>0 and φ is a continuous positive (i.e., φ>0) function in (0;ρ). If k,l=0,1,2,…(k≠l), then
∫|w|<ρwkw¯lφ(|w|)dm(w)=0,
when the corresponding integral exists.
Fact 2.
For λ≠0,-1,-2,…,
1(1-x)λ=∑k=0∞Γ(k+λ)Γ(λ)Γ(k+1)xk,|x|<1.
Fact 3.
If Rep>0,Req>0, then
∫01tp-1(1-t)q-1dt≡B(p;q)=Γ(p)Γ(q)Γ(p+q).
Fact 4.
If g∈C[0;T](T>0), then
limβ↓01Γ(β)∫0Tg(x)(T-x)β-1dx=g(T).
As a consequence of Stirling’s Formula, we have the following fact.
Fact 5.
For arbitrary μ∈C and forR≥0,R→+∞|Γ(μ+R)|≍e-R⋅RReμ+R-1/2.
In addition, if Reμ>0,Reν>0, and R≥0,R→+∞, then
|Γ(μ+R)|≍|Γ(Reμ+R)|,|Γ(μ+R)Γ(ν+R)|≍1RReν-Reμ.
Fact 6.
Assume that p>0,ρ>0, and f∈H(|w|<ρ); then ∫02π|f(reiϑ)|pdϑ is a nondecreasing function of r, that is,
∫02π|f(r1eiϑ)|pdϑ≤∫02π|f(r2eiϑ)|pdϑ,0≤r1<r2<ρ.
Corollary 2.2.
Assume that p>0,ρ>0,f∈H(|w|<ρ), and φ is a continuous positive (i.e., φ>0) function in (0;ρ). Then
∫ρ1<|w|<ρ2|f(r⋅w)|p⋅φ(|w|)dm(w)≤∫ρ1<|w|<ρ2|f(w)|p⋅φ(|w|)dm(w)
if 0≤ρ1<ρ2≤ρ and 0≤r≤1. In particular,
∫|w|<ρ|f(r⋅w)|p⋅φ(|w|)dm(w)≤∫|w|<ρ|f(w)|p⋅φ(|w|)dm(w).
3. Main Function Spaces
Suppose that α=(α1,α2,…,αn)∈Rn. We put
mα=∑αi<0-αi,lα=∑αi>0αi.
Further, we shall write α≺ (⋆) only if the following conditions are satisfied:
αn+αn-1+⋯+αi>-(n+1-i)(1≤i≤n).
Similarly, if β=(β1,β2,…,βn)∈Cn, then we shall write β≺ (⋆) if only Reβ≡(Reβ1,Reβ2,…,Reβn)≺ (⋆).
The following multidimensional analogue of Fact 1 is valid.
Lemma 3.1.
For α=(α1,α2,…,αn)≺ (⋆) and for arbitrary multi-index k=(k1,k2,…,kn)∈Z+n, put
J(n)(α;k)≡J(n)(α1,α2,…,αn;k1,k2,…,kn)=∫Bn∏i=1n|wi|2ki∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w);
then
J(n)(α;k)=πn⋅Γ(kn+αn+αn-1+2)Γ(kn+αn+2)×Γ(kn+kn-1+αn+αn-1+αn-2+3)Γ(kn+kn-1+αn+αn-1+3)×⋯×Γ(kn+kn-1+kn-2+⋯+k3+αn+αn-1+αn-2+⋯+α3+α2+n-1)Γ(kn+kn-1+kn-2+⋯+k3+αn+αn-1+αn-2+⋯+α3+n-1)×Γ(kn+kn-1+kn-2+⋯+k3+k2+αn+αn-1+αn-2+⋯+α3+α2+α1+n)Γ(kn+kn-1+kn-2+⋯+k3+k2+αn+αn-1+αn-2+⋯+α3+α2+n)×Γ(kn+1)⋅Γ(kn-1+1)⋅Γ(kn-2+1)⋅⋯⋅Γ(k3+1)⋅Γ(k2+1)⋅Γ(k1+1)⋅Γ(αn+1)Γ(kn+kn-1+kn-2+⋯+k3+k2+k1+αn+αn-1+αn-2+⋯+α3+α2+α1+n+1).
Moreover, if α≺ (⋆) and k=(k1,k2,…,kn),l=(l1,l2,…,ln) are arbitrary multi-indices such that k≠l, then
∫Bn∏i=1nwiki∏i=1nwi¯li∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)=0.
Proof.
We intend to establish (3.3) by induction in n. For n=1 we simply arrived at (2.2). Assume the validity of (3.3) for some n and proceed to the case of n+1. Note that arbitrary w∈Bn+1 can be written as w=(w′,wn+1), where w′∈Bn,|wn+1|<1-|w′|2. Consequently, for α=(α1,α2,…,αn,αn+1)≺ (⋆) and for multi-index k=(k1,k2,…,kn,kn+1)∈Z+n+1, we have
J(n+1)(α;k)≡J(n+1)(α1,α2,…,αn,αn+1;k1,k2,…,kn,kn+1)=∫Bn+1∏i=1n+1|wi|2ki∏i=1n+1(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)=∫Bn∏i=1n|wi|2ki∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αi×∫|wn+1|<1-|w′|2|wn+1|2kn+1(1-|w′|2-|wn+1|2)αn+1dm(wn+1)dm(w′).
A simple change of variable: wn+1=1-|w′|2·ζ(ζ∈D) in the inner integral gives the following recurrent relation:
J(n+1)(α1,α2,…,αn,αn+1;k1,k2,…,kn,kn+1)=πΓ(kn+1+1)Γ(αn+1+1)Γ(kn+1+αn+1+2)×J(n)(α1,α2,…,αn-1,αn+αn+1+kn+1+1;k1,k2,…,kn).
Note that (α1,α2,…,αn-1,αn+αn+1+kn+1+1)≺ (⋆) (for all kn+1∈Z) due to the condition α=(α1,α2,…,αn,αn+1)≺ (⋆). Consequently, in (3.6) we can apply (3.3) to J(n)(α1,α2,…,αn-1,αn+αn+1+kn+1+1;k1,k2,…,kn) due to our inductive assumption. As a result, we obtain (3.3) but this time for n+1. Thus, the inductive argument is completed and the formula (3.3) is established.
Now suppose that k=(k1,k2,…,kn),l=(l1,l2,…,ln)∈Z+n, and k≠l⇒ there exists i0(1≤i0≤n) such that ki0≠li0. Then we can split arbitrary w=(w1,…,wn)∈Bn in wi0̂=(w1,…i0̂…,wn)∈Cn-1 and wi0∈C so that |wi0̂|2+|wi0|2<1 and dm(w)=dm(wi0)dm(wi0̂). Hence
∫Bn∏i=1nwiki∏i=1nwi¯li∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)=∫|wi0̂|2<1w1k1⋅w1¯l1⋅…i0̂…⋅wnknwn¯ln⋅∏i=1i0-1(1-|w1|2-|w2|2-⋯-|wi|2)αi×∫|wi0|2<1-|wi0̂|2wi0ki0⋅wi0¯li0⋅∏i=i0n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(wi0)dm(wi0̂).
In view of (2.4), the inner integral in (3.7) is equal to 0, so (3.4) is also proved.
Corollary 3.2.
If α≺ (⋆), then
∫Bn∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)=πn∏i=1n(αn+αn-1+⋯+αi+n+1-i)<+∞.
Remark 3.3.
In the integrals J(n)(α;k) (see (3.2)) instead of α≺ (⋆), arbitrary β≺ (⋆) (β∈Cn) can be considered and the formulas (3.3), (3.4), and (3.8) remain true after the replacement of α by β.
Definition 3.4.
Assume that 0<p<+∞ and α=(α1,α2,…,αn)≺ (⋆). Denote by Lαp(Bn) the space of all complex-valued functions f in Bn with
Mαp(f)≡∫Bn|f(w)|p⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)<+∞.
We obviously have
Mαp(f+g)≤2p{Mαp(f)+Mαp(g)},Mαp(c⋅f)=cp⋅Mαp(f).
Correspondingly, denote by Hαp(Bn)⊂Lαp(Bn) the subspace of functions holomorphic in Bn. Note that 1≤p<q<+∞⇒Lαq(Bn)⊂Lαp(Bn),Hαq(Bn)⊂Hαp(Bn).
Lemma 3.5.
Assume that 0<p<+∞,α≺ (⋆), and f∈H(Bn). Then
|f(z)|p≤22n+lαcn⋅(1-|z|)2n+lα⋅Mαp(f),∀z∈Bn,
where cn=m(Bn) is the volume of the unit ball of Cn.
Proof.
Fix an arbitrary z∈Bn; put r=(1-|z|)/2 and B(z;r)={w∈Cn:|w-z|<r}⊂Bn. Since |f|p is subharmonic in Bn, we have
|f(z)|p≤1cn⋅r2n∫B(z;r)|f(w)|pdm(w).
Note that
w∈B(z;r)⟹|w|<r+|z|=1-|z|2+|z|=1+|z|2⟹1-|w|>1-|z|2⟹1-|w|2≥1-|w|>1-|z|2⟹1-|w1|2-|w2|2-⋯-|wi|2>1-|z|2(1≤i≤n).
Hence for w∈B(z;r)(1-|w1|2-|w2|2-⋯-|wi|2)αi≥{1,αi≤0,(1-|z|)αi2αi,αi>0.
Combining (3.12) and (3.14), we obtain
|f(z)|p≤1cn⋅((1-|z|)/2)2n⋅∏αi>0((1-|z|)αi/2αi)∏αi>0((1-|z|)αi/2αi)∫B(z;r)|f(w)|pdm(w)≤22n+lαcn⋅(1-|z|)2n+lα∫B(z;r)|f(w)|p⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)≤22n+lαcn⋅(1-|z|)2n+lα⋅Mαp(f).
Corollary 3.6.
Hαp(Bn) is a closed subspace in Lαp(Bn).
Definition 3.7.
If f∈H(Bn) and 0≤ri≤1(1≤i≤n), then put
fr1r2⋯rn(w)≡f(r1⋅w1,r2⋅w2,…,rn⋅wn),w=(w1,w2,…,wn)∈Bn.
Similarly, if f∈H(Bn) and 0≤r≤1, then put
fr(w)≡f(r⋅w1,r⋅w2,…,r⋅wn),w=(w1,w2,…,wn)∈Bn.
In particular, if ri≡r∈[0;1](1≤i≤n)⇒fr1r2⋯rn(w)≡fr(w), note also that f11⋯1(w)≡f1(w)≡f(w).
Lemma 3.8.
Assume that 0<p<+∞,α≺ (⋆), and f∈H(Bn). Then for all ri,0≤ri≤1(1≤i≤n),
Mαp(fr1r2⋯rn)≤Mαp(f).
In particular, f∈Hαp(Bn)⇒fr1r2⋯rn∈Hαp(Bn).
Proof.
Evidently, it suffices to fix i0(1≤i0≤n) and consider the case 0≤ri0≤1,ri=1(1≤i≤n,i≠i0). In other words, it is sufficient to show that
Mαp(f11⋯1ri01⋯11)≤Mαp(f).
To this end, we proceed as follows:
Mαp(f11⋯1ri01⋯11)=∫Bn|f(w1,…,wi0-1,ri0⋅wi0,wi0+1,…,wn)|p×∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)=∫|wi0̂|2<1∏i=1i0-1(1-|w1|2-|w2|2-⋯-|wi|2)αi×∫|wi0|2<1-|wi0̂|2|f(w1,…,wi0-1,ri0⋅wi0,wi0+1,…,wn)|p×∏i=i0n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(wi0)dm(wi0̂).
An application of Corollary 2.2 (see (2.12)) to the inner integral gives the desired inequality.
Corollary 3.9.
Assume that 0<p<+∞,α≺ (⋆), and f∈H(Bn). Then for all r,0≤r≤1,
Mαp(fr)≤Mαp(f).
In particular, f∈Hαp(Bn)⇒fr∈Hαp(Bn).
Lemma 3.10.
Assume that 0<p<+∞,α≺ (⋆), and f∈H(Bn). Then for all ri,0≤ri≤1(1≤i≤n) and for arbitrary ε(0<ε≤1),
∫1-ε<|w|<1|fr1r2⋯rn(w)|p⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)≤∫1-ε<|w|<1|f(w)|p⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w).
Proof.
Since the case ε=1 coincides with (3.18), we can suppose that 0<ε<1. Further, similarly to the proof of Lemma 3.8, it suffices to fix i0(1≤i0≤n) and consider the case 0≤ri0≤1,ri=1(1≤i≤n,i≠i0):∫1-ε<|w|<1|f(w1,…,wi0-1,ri0⋅wi0,wi0+1,…,wn)|p×∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)≡I1+I2,
where
I1=∫|wi0̂|2<(1-ε)2∏i=1i0-1(1-|w1|2-|w2|2-⋯-|wi|2)αi×∫(1-ε)2-|wi0̂|2<|wi0|2<1-|wi0̂|2|f(w1,…,wi0-1,ri0⋅wi0,wi0+1,…,wn)|p×∏i=i0n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(wi0)dm(wi0̂),I2=∫(1-ε)2<|wi0̂|2<1∏i=1i0-1(1-|w1|2-|w2|2-⋯-|wi|2)αi×∫|wi0|2<1-|wi0̂|2|f(w1,…,wi0-1,ri0wi0,wi0+1,…,wn)|p×∏i=i0n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(wi0)dm(wi0̂).
It remains to apply Corollary 2.2. More exactly, let us apply (2.11) to the inner integral of (3.24) and (2.12) to the inner integral of (3.25). By this, the proof is complete.
Corollary 3.11.
Assume that 0<p<+∞,α≺ (⋆), and f∈H(Bn). Then for all r,0≤r≤1 and for arbitrary ε(0<ε≤1),
∫1-ε<|w|<1|fr(w)|p⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)≤∫1-ε<|w|<1|f(w)|p⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w).
Lemma 3.12.
Assume that 0<p<+∞,α≺ (⋆), and f∈Hαp(Bn). Then
Mαp(f-fr1r2⋯rn)⟶0asri↑1(1≤i≤n).
Proof.
For arbitrary fixed ε∈(0;1) and for all ri,0≤ri≤1(1≤i≤n), we have
Mαp(f-fr1r2⋯rn)=∫Bn|f(w1,w2,…,wn)-f(r1⋅w1,r2⋅w2,…,rn⋅wn)|p×∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)=∫|w|<1-ε+∫1-ε<|w|<1≡J1+J2.
In view of the inequality (a+b)p≤2p(ap+bp)(a,b≥0,p>0) and due to (3.22),
J2≤2p+1⋅∫1-ε<|w|<1|f(w1,w2,…,wn)|p×∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w).
Note that the condition f∈Hαp(Bn) implies J2→0 as ε↓0. Besides, it is evident that for a fixed ε∈(0;1):J1→0 as ri↑1(1≤i≤n). All these complete the proof.
Corollary 3.13.
Assume that 0<p<+∞,α≺ (⋆), and f∈Hαp(Bn). Then
Mαp(f-fr)⟶0asr↑1.
4. Main Integral Representations
In fact, Lemma 3.1 asserts that the system{w1k1⋅w2k2⋅⋯⋅wnkn}k1,k2,…,kn=0∞
is orthogonal in the Hilbert space L2{Bn;∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)}.
And what about completeness of this system?
Proposition 4.1.
Assume that 1≤p<+∞,α≺ (⋆), and f∈Hαp(Bn). If for arbitrary multi-index k=(k1,k2,…,kn)∈Z+n∫Bnf(w)⋅w1¯k1⋅w2¯k2⋅⋯⋅wn¯kn⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)=0,
then f≡0.
Proof.
Since f∈H(Bn)⇒f(w)=∑l1,l2,…,ln≥0al1l2⋯ln⋅w1l1w2l2⋯wnln≡∑|l|≥0al⋅wl,w∈Bn,
where |l|=l1+l2+⋯+ln for arbitrary multi-index l=(l1,l2,…,ln)∈Z+n. Moreover, for arbitrary compact K⊂Bn there exist a positive series ∑|l|≥0cl<+∞ such that |al·wl|≤cl(|l|≥0) uniformly in w∈K. Consequently, for arbitrary fixed r∈(0;1)f(r⋅w)=∑|l|≥0al⋅r|l|wl,w∈Bn¯,
and, in addition, there exist a positive series ∑|l|≥0dl<+∞ such that |al·r|l|·wl|≤dl(|l|≥0) uniformly in w∈Bn¯.
Now let us fix arbitrary k1≥0,k2≥0,…,kn≥0. Since fr→f (as r↑1) in the space Hαp(Bn) (Corollary 3.13), we have (as r↑1)
∫Bnf(r⋅w)⋅∏i=1nwi¯ki⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)⟶∫Bnf(w)⋅∏i=1nwi¯ki⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)=0.
On the other hand,
∫Bnf(r⋅w)⋅∏i=1nwi¯ki⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)=∑|l|≥0al⋅r|l|∫Bn∏i=1nwili⋅∏i=1nwi¯ki⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)=ak⋅r|k|⋅J(n)(α1,…,αn;k1,…,kn).
Combining (4.6) and (4.7), we obtain that ak·r|k|·J(n)(α1,…,αn;k1,…,kn)→0 (as r↑1) ⇒ak=0 and this takes place for any multi-index k⇒f≡0.
Corollary 4.2.
If α≺ (⋆), then the system
{w1k1⋅w2k2⋅⋯⋅wnknJ(n)(α1,…,αn;k1,…,kn)}k1,k2,…,kn=0∞
is an orthonormal basis in the space Hα2(Bn).
Corollary 4.3.
If α≺ (⋆), then the reproducing kernel (i.e., the kern-function) of the space Hα2(Bn) is given by the formula
Sα(z;w)=∑k1,k2,…,kn=0∞(z1w1¯)k1⋅(z2w2¯)k2⋅⋯⋅(znwn¯)knJ(n)(α1,…,αn;k1,…,kn),z,w∈Bn.
Definition 4.4.
Assume that β=(β1,β2,…,βn)∈Cn and β≺ (⋆) ⇔Reβ=(Reβ1,Reβ2,…,Reβn)≺ (⋆). For arbitrary z∈Bn,w∈Bn¯, let us introduce the following kernel (motivated by (4.9)):
Sβ(z;w)=∑k1,k2,…,kn=0∞(z1w1¯)k1⋅(z2w2¯)k2⋅⋯⋅(znwn¯)knJ(n)(β1,…,βn;k1,…,kn)
or, in a slightly different form,
Sβ(z;w)=∑k=0∞∑k1+k2+⋯+kn-1+kn=k(z1w1¯)k1⋅(z2w2¯)k2⋅⋯⋅(znwn¯)knJ(n)(β1,…,βn;k1,…,kn)≡∑k=0∞Pk(z;w),
where
Pk(z;w)=1πn⋅∑k1+k2+⋯+kn-1+kn=k(z1w1¯)k1⋅(z2w2¯)k2⋅⋯⋅(znwn¯)kn×Γ(kn+βn+2)Γ(kn+βn+βn-1+2)×Γ(kn+kn-1+βn+βn-1+3)Γ(kn+kn-1+βn+βn-1+βn-2+3)×⋯×Γ(kn+kn-1+kn-2+⋯+k3+βn+βn-1+βn-2+⋯+β3+n-1)Γ(kn+kn-1+kn-2+⋯+k3+βn+βn-1+βn-2+⋯+β3+β2+n-1)×Γ(kn+kn-1+kn-2+⋯+k3+k2+βn+βn-1+βn-2+⋯+β3+β2+n)Γ(kn+kn-1+kn-2+⋯+k3+k2+βn+βn-1+βn-2+⋯+β3+β2+β1+n)×Γ(kn+kn-1+kn-2+⋯+k3+k2+k1+βn+βn-1+βn-2+⋯+β3+β2+β1+n+1)Γ(kn+1)⋅Γ(kn-1+1)⋅Γ(kn-2+1)⋅⋯⋅Γ(k3+1)⋅Γ(k2+1)⋅Γ(k1+1)⋅Γ(βn+1).
Note that if we take Reβn>-1,βi=0(1≤i≤n-1) in (4.9’’)-(4.10),then
Sβ(z;w)=Γ(βn+n+1)πn⋅Γ(βn+1)⋅1(1-z1w1¯-z2w2¯-z3w3¯-⋯-znwn¯)βn+n+1,
that is, we arrive at the kernel of the integral representations (1.4).
We are going to estimate Pk(z;w),k=0,1,2,…⇒ to estimate the kernel Sβ(z;w). Let us put (compare with (3.1))
mβ⋆=∑Reβi<0,1≤i≤n-1(-Reβi),lβ⋆=∑Reβi>0,1≤i≤n-1(Reβi).
Then in view of (2.9), we have (z∈Bn,w∈Bn¯,k=0,1,2,…)
|Pk(z;w)|≤const(n;β)⋅∑k1+k2+⋯+kn-1+kn=k1(kn+1)Reβn-1×1(kn+kn-1+1)Reβn-2×⋯×1(kn+kn-1+kn-2+⋯+k3+1)Reβ2×1(kn+kn-1+kn-2+⋯+k3+k2+1)Reβ1×|Γ(kn+kn-1+⋯+k2+k1+βn+βn-1+⋯+β2+β1+n+1)||Γ(βn+1)|⋅Γ(kn+kn-1+⋯+k2+k1+1)×[Γ(kn+kn-1+⋯+k2+k1+1)⋅|z1w1¯|k1⋅|z2w2¯|k2⋅⋯⋅|znwn¯|knΓ(kn+1)⋅Γ(kn-1+1)⋅⋯⋅Γ(k2+1)⋅Γ(k1+1)]≤const(n;β)⋅∑k1+k2+⋯+kn-1+kn=k(kn+kn-1+kn-2+⋯+k3+k2+k1+1)mβ⋆(kn+1)lβ⋆×|Γ(kn+kn-1+⋯+k2+k1+Reβn+Reβn-1+⋯+Reβ2+Reβ1+n+1)||Γ(Reβn+1)|⋅Γ(kn+kn-1+⋯+k2+k1+1)×[Γ(kn+kn-1+⋯+k2+k1+1)⋅|z1w1¯|k1⋅|z2w2¯|k2⋅⋯⋅|znwn¯|knΓ(kn+1)⋅Γ(kn-1+1)⋅⋯⋅Γ(k2+1)⋅Γ(k1+1)]≤const(n;β)⋅(k+1)mβ⋆⋅Γ(k+Reβn+Reβn-1+⋯+Reβ2+Reβ1+n+1)Γ(Reβn+1)⋅Γ(k+1)×∑k1+k2+⋯+kn-1+kn=kΓ(k+1)⋅|z1w1¯|k1⋅|z2w2¯|k2⋅⋯⋅|znwn¯|knΓ(kn+1)⋅Γ(kn-1+1)⋅⋯⋅Γ(k2+1)⋅Γ(k1+1)≤const(n;β)⋅(k+1)mβ⋆⋅Γ(k+Reβn+Reβn-1+⋯+Reβ2+Reβ1+n+1)Γ(Reβn+Reβn-1+⋯+Reβ2+Reβ1+n+1)⋅Γ(k+1)×(|z1w1¯|+|z2w2¯|+⋯+|znwn¯|)k.
Further, due to (2.8),
Γ(k+Reβn+Reβn-1+⋯+Reβ2+Reβ1+n+1)≍e-(k+1)⋅(k+1)Reβn+Reβn-1+⋯+Reβ2+Reβ1+n+k+1-1/2(k=0,1,2,…).
Consequently,
(k+1)mβ⋆⋅Γ(k+Reβn+Reβn-1+⋯+Reβ2+Reβ1+n+1)≍e-(k+1)⋅(k+1)Reβn+Reβn-1+⋯+Reβ2+Reβ1+n+mβ⋆+k+1-1/2≍Γ(k+Reβn+Reβn-1+⋯+Reβ2+Reβ1+mβ⋆+n+1).
Combining (4.13), (4.15) and taking into account the inequality
|z1w1¯|+|z2w2¯|+⋯+|znwn¯|≤|z|⋅|w|≤|z|<1(z∈Bn,w∈Bn¯),
we obtain
|Pk(z;w)|≤const(n;β)⋅Γ(k+Reβn+Reβn-1+⋯+Reβ2+Reβ1+mβ⋆+n+1)⋅|z|kΓ(Reβn+Reβn-1+⋯+Reβ2+Reβ1+mβ⋆+n+1)⋅Γ(k+1).
In view of the formula (2.5), the summation of (4.17) over k=0,1,2,… gives
|Sβ(z;w)|≤const(n;β)(1-|z|)Reβn+Reβn-1+⋯+Reβ2+Reβ1+mβ⋆+n+1=const(n;β)(1-|z|)Reβn+lβ⋆+n+1.
Thus, we have the following theorem.
Theorem 4.5.
If β≺ (⋆), then the kernel Sβ(z;w) satisfies the following properties (z∈Bn,w∈Bn¯):
the series (4.9’) converges absolutely for arbitrary fixed z∈Bn,w∈Bn¯;
in the expansion Sβ(z;w)=∑k=0∞Pk(z;w) (see (4.9’’)), the series is majorated by a convergent positive numerical series ∑k=0∞bk uniformly in z∈K⊂⊂Bn,w∈Bn¯;
Sβ(z;w) is holomorphic in z∈Bn;
Sβ(z;w) is antiholomorphic in w∈Bn and continuous in w∈Bn¯;
the estimate
|Sβ(z;w)|≤const(n;β)(1-|z|)n+1+Reβn+lβ⋆
is valid uniformly in z∈Bn,w∈Bn¯.
Definition 4.6.
Assume that 1≤p<+∞ and α=(α1,α2,…,αn)≺ (⋆). For β=(β1,β2,…,βn)∈Cn, we shall write β≻(p;α) if only (compare with (1.3))
when p=1, then
Reβi≥αi(1≤i≤n);
when 1<p<+∞, then
Reβn+Reβn-1+⋯+Reβi>αn+αn-1+⋯+αi+n+1-ip-(n+1-i).
Obviously, β≻(p;α)⇒β≺ (⋆).
Theorem 4.7.
Assume that 1≤p<+∞,α≺ (⋆), and β≻(p;α). Then arbitrary f∈Hαp(Bn) admits the following integral representations:
f(z)=∫Bnf(w)Sβ(z;w)∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)βidm(w),z∈Bn,f(0)¯=∫Bnf(w)¯Sβ(z;w)∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)βidm(w),z∈Bn.
Proof.
Let us fix an arbitrary f∈Hαp(Bn) and an arbitrary z∈Bn. Further,
f(w)=∑|l|≥0al⋅wl=∑m=0∞Fm(w),w∈Bn,
where
Fm(w)=∑l1+l2+⋯+ln-1+ln=mal⋅wl(m=0,1,2,…).
Hence, for arbitrary fixed r∈(0;1),
fr(w)=∑|l|≥0al⋅r|l|⋅wl=∑m=0∞rm⋅Fm(w),w∈Bn¯.
Obviously,
fr(w)¯=∑|l|≥0al¯⋅r|l|⋅w¯l=∑m=0∞rm⋅Fm(w)¯,w∈Bn¯.
Note that the first series in (4.26) and (4.27) and the second series in (4.26) and (4.27) are majorated (uniformly in w∈Bn¯) by positive convergent numerical series ∑|l|≥0cl<+∞ and ∑m=0∞cm̃<+∞, respectively.
First we intend to prove both formulas (4.22), (4.23) for the function fr, that is,
fr(z)=∫Bnfr(w)Sβ(z;w)∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)βidm(w),f(0)¯=∫Bnfr(w)¯Sβ(z;w)∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)βidm(w).
In view of Theorem 4.5(b), (4.26)-(4.27) and due to Lebesgue’s dominated convergence theorem, we have
∫Bnfr(w)Sβ(z;w)∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)βidm(w)=∑k,m=0∞∫BnrmFm(w)⋅Pk(z;w)⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)βidm(w),∫Bnfr(w)¯Sβ(z;w)∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)βidm(w)=∑k,m=0∞∫BnrmFm(w)¯⋅Pk(z;w)⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)βidm(w),
and the series ∑k,m=0∞… in (4.28) and (4.29) converge absolutely.
Taking into account (3.4) and Remark 3.3, we conclude that all terms in the right-hand side of (4.29) (except of the case k=m=0) are equal to zero. Hence the left-hand side of (4.29) is equal to
∫Bna00⋯0¯⋅1πn⋅Γ(βn+2)Γ(βn+βn-1+2)×Γ(βn+βn-1+3)Γ(βn+βn-1+βn-2+3)×⋯×Γ(βn+βn-1+βn-2+⋯+β3+n-1)Γ(βn+βn-1+βn-2+⋯+β3+β2+n-1)×Γ(βn+βn-1+βn-2+⋯+β3+β2+n)Γ(βn+βn-1+βn-2+⋯+β3+β2+β1+n)×Γ(βn+βn-1+βn-2+⋯+β3+β2+β1+n+1)Γ(βn+1)×∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)βidm(w)=f(0)¯⋅∫Bn∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)βidm(w)J(n)(β1,β2,…,βn;0,0,…,0)=f(0)¯.
Thus, (4.23’) has been proved.
Now let us proceed to (4.22’). If k≠m, then in view of (3.4) and Remark 3.3∫BnrmFm(w)⋅Pk(z;w)⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)βidm(w)=0.
If k=m, then in view of (3.4), (3.2), (3.3) and Remark 3.3∫BnrkFk(w)⋅Pk(z;w)⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)βidm(w)=rkπn⋅∑k1+k2+⋯+kn-1+kn=kak1k2⋯kn-1knz1k1⋅z2k2⋅⋯⋅zn-1kn-1⋅znknΓ(kn+1)⋅Γ(kn-1+1)⋅⋯⋅Γ(k2+1)⋅Γ(k1+1)×Γ(kn+βn+2)Γ(kn+βn+βn-1+2)×Γ(kn+kn-1+βn+βn-1+3)Γ(kn+kn-1+βn+βn-1+βn-2+3)×⋯×Γ(kn+kn-1+kn-2+⋯+k3+βn+βn-1+βn-2+⋯+β3+n-1)Γ(kn+kn-1+kn-2+⋯+k3+βn+βn-1+βn-2+⋯+β3+β2+n-1)×Γ(kn+kn-1+kn-2+⋯+k3+k2+βn+βn-1+βn-2+⋯+β3+β2+n)Γ(kn+kn-1+kn-2+⋯+k3+k2+βn+βn-1+βn-2+⋯+β3+β2+β1+n)×Γ(kn+kn-1+kn-2+⋯+k3+k2+k1+βn+βn-1+βn-2+⋯+β3+β2+β1+n+1)Γ(βn+1)×J(n)(β1,β2,…,βn;k1,k2,…,kn)=rk⋅∑k1+k2+⋯+kn-1+kn=kak1k2⋯kn-1knz1k1⋅z2k2⋅⋯⋅zn-1kn-1⋅znkn⋅1=rk⋅Fk(z).
Consequently, (4.28), (4.31), and (4.32) together yield
∫Bnfr(w)Sβ(z;w)∏k=1n(1-|w1|2-|w2|2-⋯-|wk|2)βkdm(w)=∑k=0∞rk⋅Fk(z)=fr(z).
Thus, (4.22’) also has been proved.
Now we intend to make passage r↑1 in (4.22’) and (4.23’). Evidently, the left-hand sides tend to f(z) and f(0)¯ correspondingly. It remains to show that the right-hand sides of (4.22’) and (4.23’) tend to the right-hand sides of (4.22) and (4.23), respectively. In view of the estimate (4.19), it suffices to show that
I(r)≡∫Bn|f(w)-fr(w)|⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)Reβidm(w)⟶r↑10.
If p=1, then due to the condition (4.20) and Corollary 3.13I(r)=∫Bn|f(w)-fr(w)|⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)Reβi-αi+αidm(w)≤∫Bn|f(w)-fr(w)|⋅∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)αidm(w)=Mα1(f-fr)⟶r↑10.
If 1<p<∞, then an application of Hölder integral inequality to I(r) gives
I(r)≤[Mαp(f-fr)]1/p⋅(∫Bn∏i=1n(1-|w1|2-|w2|2-⋯-|wi|2)q(Reβi-αi/p)dm(w))1/q.
Here the integral over Bn is finite due to Corollary 3.2 and the condition β≻(p;α). Consequently, I(r)→0,r↑1, in view of Corollary 3.13. Thus, the theorem is proved.
Theorem 4.8.
For Reβn>-1,Reβi>0(1≤i≤n-1), and for z∈Bn,w∈Bn¯,
Sβ(z;w)=Γ(βn+βn-1+⋯+β2+β1+n+1)πn⋅Γ(β1)Γ(β2)⋅⋯⋅Γ(βn-1)Γ(βn+1)×∫[0;1]n-1∏i=n2siβn+βn-1+⋯+βi+n+1-i⋅∏i=n2(1-si)βi-1-1dsndsn-1⋯ds3ds2(1-z1w1¯-s2z2w2¯-s2s3z3w3¯-⋯-s2s3⋯snznwn¯)βn+βn-1+⋯+β2+β1+n+1.
Proof.
For k=0,1,2,3,… according to (4.10) and (2.6), we have
Pk(z;w)=1πn⋅∑k1+k2+⋯+kn-1+kn=k(z1w1¯)k1⋅(z2w2¯)k2⋅⋯⋅(zn-1wn-1¯)kn-1⋅(znwn¯)knΓ(k1+1)⋅Γ(k2+1)⋅⋯⋅Γ(kn-1+1)⋅Γ(kn+1)×1Γ(βn-1)∫01snkn+βn+1(1-sn)βn-1-1dsn×1Γ(βn-2)∫01sn-1kn+kn-1+βn+βn-1+2(1-sn-1)βn-2-1dsn-1×⋯×1Γ(β1)∫01s2kn+kn-1+kn-2+⋯+k2+βn+βn-1+βn-2+⋯+β2+n-1(1-s2)β1-1ds2×Γ(kn+kn-1+kn-2+⋯+k2+k1+βn+βn-1+βn-2+⋯+β2+β1+n+1)Γ(βn+1)=1πn⋅Γ(β1)Γ(β2)⋯Γ(βn-1)Γ(βn+1)×∫[0;1]n-1snβn+1⋅sn-1βn+βn-1+2⋅⋯⋅s2βn+βn-1+⋯+β2+n-1×(1-sn)βn-1-1⋅(1-sn-1)βn-2-1⋅⋯⋅(1-s2)β1-1×∑k1+k2+⋯+kn-1+kn=k(z1w1¯)k1⋅(s2z2w2¯)k2⋅(s2s3z3w3¯)k3⋅⋯⋅(s2s3⋯snznwn¯)knΓ(kn+1)⋅Γ(kn-1+1)⋅⋯⋅Γ(k2+1)⋅Γ(k1+1)×Γ(k+1)Γ(k+1)⋅Γ(k+βn+βn-1+⋯+β2+β1+n+1)dsndsn-1⋯ds3ds2=Γ(k+βn+βn-1+⋯+β2+β1+n+1)πn⋅Γ(k+1)Γ(β1)Γ(β2)⋯Γ(βn-1)Γ(βn+1)×∫[0;1]n-1snβn+1⋅sn-1βn+βn-1+2⋅⋯⋅s2βn+βn-1+⋯+β2+n-1×(1-sn)βn-1-1⋅(1-sn-1)βn-2-1⋅⋯⋅(1-s2)β1-1×(z1w1¯+s2z2w2¯+s2s3z3w3¯+⋯+s2s3⋯snznwn¯)kdsndsn-1⋯ds3ds2=Pk(z;w).
The summation of these relations over k=0,1,2,3,… yields (see (2.5))
Sβ(z;w)=∑k=0∞Pk(z;w)=Γ(βn+βn-1+βn-2+⋯+β2+β1+n+1)πn⋅Γ(β1)Γ(β2)⋯Γ(βn-1)Γ(βn+1)×∫[0;1]n-1snβn+1⋅sn-1βn+βn-1+2⋅⋯⋅s2βn+βn-1+⋯+β2+n-1×(1-sn)βn-1-1⋅(1-sn-1)βn-2-1⋅⋯⋅(1-s2)β1-1×dsndsn-1⋯ds3ds2(1-z1w1¯-s2z2w2¯-s2s3z3w3¯-⋯-s2s3⋯snznwn¯)βn+βn-1+⋯+β2+β1+n+1.
Thus, (4.37) is proved.
Remark 4.9.
Under the conditions of the theorem, the formula (4.37) easily implies the assertions (c), (d), and (e) of Theorem 4.5.
Remark 4.10.
If we take Reβn>-1,βi↓0(1≤i≤n-1) in (4.37), then the formal application of (2.7) gives the following formula:
Sβ(z;w)=Γ(βn+n+1)πn⋅Γ(βn+1)⋅1(1-z1w1¯-z2w2¯-z3w3¯-⋯-znwn¯)βn+n+1,
that is, we arrive at the kernel of the integral representations (1.4).
Remark 4.11.
In fact, the kernel Sβ(z;w), defined for β≺ (⋆) by (4.9’)-(4.9’’), can be considered as an analytic continuation in β of the integral Sβ(z;w), defined for Reβn>-1,Reβi>0(1≤i≤n-1) by (4.37).
Now we consider an interesting special case n=2, when (in comparison with (4.9’)-(4.9’’) or (4.37)) the kernel Sβ(z;w) can be written out in an explicit form.
Theorem 4.12.
Assume that β=(β1,β2)≺ (⋆), that is, Reβ2>-1,Reβ2+Reβ1>-2. Then for z=(z1,z2)∈B2,w=(w1,w2)∈B2¯,
Sβ(z;w)=1π2⋅β1(β2+1)(1-z1w1¯)β1+1(1-z1w1¯-z2w2¯)β2+2+1π2⋅(β2+1)(β2+2)(1-z1w1¯)β1(1-z1w1¯-z2w2¯)β2+3.
Proof.
According to (4.9’)-(4.9’’)-(4.10) and in view of the formula (2.5),
Sβ(z;w)=1π2⋅∑k1,k2=0∞(z1w1¯)k1⋅(z2w2¯)k2Γ(k1+1)Γ(k2+1)Γ(β2+k2+2)Γ(β2+β1+k2+2)⋅Γ(β2+β1+k2+k1+3)Γ(β2+1)=1π2⋅∑k2=0∞(z2w2¯)k2Γ(k2+1)⋅Γ(β2+k2+2)Γ(β2+β1+k2+2)⋅Γ(β2+1)×∑k1=0∞(z1w1¯)k1Γ(k1+1)⋅Γ(β2+β1+k2+k1+3)=1π2⋅∑k2=0∞(z2w2¯)k2Γ(k2+1)⋅Γ(β2+k2+2)⋅Γ(β2+β1+k2+3)Γ(β2+β1+k2+2)⋅Γ(β2+1)1(1-z1w1¯)β2+β1+k2+3=1π2⋅1(1-z1w1¯)β2+β1+3⋅∑k2=0∞(z2w2¯1-z1w1¯)k2Γ(β2+k2+2)Γ(k2+1)Γ(β2+1)⋅(β2+β1+k2+2)=1π2⋅1(1-z1w1¯)β2+β1+3⋅∑k2=0∞(z2w2¯1-z1w1¯)k2×{β1Γ(β2+k2+2)Γ(k2+1)Γ(β2+1)+Γ(β2+k2+3)Γ(k2+1)Γ(β2+1)}=1π2⋅1(1-z1w1¯)β2+β1+3{β1(β2+1)(1-z2w2¯/(1-z1w1¯))β2+2+(β2+1)(β2+2)(1-z2w2¯/(1-z1w1¯))β2+3}=1π2⋅1(1-z1w1¯)β2+β1+3{β1(β2+1)(1-z1w1¯)β2+2(1-z1w1¯-z2w2¯)β2+2+(β2+1)(β2+2)(1-z1w1¯)β2+3(1-z1w1¯-z2w2¯)β2+3}=1π2⋅{β1(β2+1)(1-z1w1¯)β1+1(1-z1w1¯-z2w2¯)β2+2+(β2+1)(β2+2)(1-z1w1¯)β1(1-z1w1¯-z2w2¯)β2+3}.
Thus, (4.41) is established.
Remark 4.13.
During the proof we regroup the double series: ∑k1,k2=0∞{…}=∑k2=0∞{∑k1=0∞{…}}, which is legitimate in view of Theorem 4.5(a). In fact, we apply Fubini’s theorem for double series.
Remark 4.14.
The analysis of the proof shows that we have established (4.41) only for those z=(z1,z2)∈B2 and w=(w1,w2)∈B2¯, which satisfy the condition |z2w2¯/(1-z1w1¯)|<1. In fact, this is quite sufficiently since both sides of (4.41) are holomorphic in z∈Bn, antiholomorphic in w∈Bn, and continuous in w∈Bn¯ (see Theorem 4.5(c), (d)).
Remark 4.15.
If one takes Reβ2>-1,β1=0 in Theorem 4.12, then
Sβ1;β2(z1,z2;w1,w2)=1π2⋅(β2+1)(β2+2)(1-z1w1¯-z2w2¯)β2+3,
which coincides with (4.11) (or (4.40)) for the case n=2.
Remark 4.16.
Note that for the same case n=2 and under slightly restrictive conditions Reβ2>-1,Reβ1>0, the formula (4.37) gives (z∈B2,w∈B2¯)
Sβ(z;w)=Γ(β2+β1+3)π2⋅Γ(β2+1)Γ(β1)⋅∫01sβ2+1⋅(1-s)β1-1(1-z1w1¯-s⋅z2w2¯)β2+β1+3ds.
Remark 4.17.
It can be shown (we omit the proof) that under the conditions Reβ2>0,Reβ2+Reβ1>-2 the following interesting formula is valid (z∈B2,w∈B2¯):
S(β1,β2)(z1,z2;w1,w2)={z2w2¯⋅∂/∂(z2w2¯)+β2+1β2}⋅S(β1+1,β2-1)(z1,z2;w1,w2).
DjrbashianM. M.On the representability of certain classes of functions meromorphic in the unit disc19453139DjrbashianM. M.On the problem of representability of analytic functions19482340ForelliF.RudinW.Projections on spaces of holomorphic functions in balls1974/7524593602035786610.1512/iumj.1974.24.24044DjrbashianM. M.Survey of some achievements of Armenian mathematicians in the theory of integral representations and factorization of analytic functions1987393263282935684DjrbashianM. M.A brief survey of the results obtained by Armenian mathematicians in the field of factorization theory of meromorphic functions and its applications1988236834DjrbashianA. E.ShamoyanF. A.1988105Leipzig, GermanyBSB B. G. Teubner Verlagsgesellschaft199Teubner-Texte zur Mathematik1021691ZBL0667.30032RudinW.1980241New York, NY, USASpringerxiii+436Grundlehren der Mathematischen Wissenschaften601594