• Nebyly nalezeny žádné výsledky

Correction to "Uniformization of K/ihler manifolds with vanishing Bochner tensor"

N/A
N/A
Protected

Academic year: 2022

Podíl "Correction to "Uniformization of K/ihler manifolds with vanishing Bochner tensor""

Copied!
2
0
0

Načítání.... (zobrazit plný text nyní)

Fulltext

(1)

Acta Math., 195 (2005), 265 266

@ 2005 by Institut Mittag-Leffler. All rights reserved

Correction to

"Uniformization of K/ihler manifolds with vanishing Bochner tensor"

b y

YOSHINOBU KAMISHIMA Tokyo Metropolitan University

Tokyo, Japan

The article appeared in Acta Math., 172 (1994), 299-308

In [2], we discussed the uniformization of K/ihler manifolds with vanishing Bochner tensor, called Bochner K/ihler or Bochner-flat manifolds. T h e resulting uniformization theorem, Theorem A, was used to classify compact Bochner-flat Kfihler manifolds. In Theorem A, we claimed t h a t every Bochner-flat Kiihler manifold is uniformized by one of four types of Hermitian symmetric space. Subsequent work by R. Bryant [1] has revealed that this statement is false and that there are many Bochner-flat Kghler manifolds which are not locally symmetric. In the compact case, however, the statement of our classification result turned out to be correct, and a proof was given in [1].

In order to explain the error and to indicate how it might be corrected, let us recall the argument. The main idea, due to Webster [3], is to observe that over any simply- connected domain U in a K/ihler 2n-manifold M, there is a CR structure on p: U x R--+M whose contact form w satisfies d a b = p ' f t, where f~ is the K/~hler form of M. The contact distribution ker a~ is transverse to the fibres of p and is equipped with the lift of the complex structure on TM. There is a natural fibrewise R-action by CR automorphisms.

If M is Bochner-flat, it follows from [3] that this CR structure is spherical, and therefore there is a developing map dev: U x R - - ~ S 2n+1, together with an induced group homomorphism 0: R - - + P U ( n + I , 1) into the group of CR automorphisms of S 2n+1. This pair is uniquely determined up to CR automorphism. We let G denote the closure of 0(R) and X the complement of its fixed-point set in $2n+1: since the natural fibrewise R-action is free and dev is an immersion, it follows that d e v ( U x R ) c X .

If we have a good open cover Us of M, we have developing pairs (dev~, O~) related by CR automorphisms on pairwise intersections, and (assuming t h a t M is connected),

(2)

2 6 6 Y. K A M I S H I M A

after conjugating by CR automorphisms, we can suppose t h a t 0 : = 0 ~ coincide for all a, hence also G and X are independent of c~.

A uniformization result, would then follow by passing to the quotient of X by ~(R).

In [2, w we implicitly assumed t h a t the quotient g e o m e t r y was determined uniquely by X (also we o m i t t e d the case X = S 2 ~ + I - { 0 , at}), and so we did not take full account of the possible conjugacy classes of 0(R). Furthermore in the case t h a t Q(R) is not closed, the quotient of X by 0(R) is not a manifold, so we need to work with local quotients.

W h e n M is compact, the non-closed case does not occur: establishing this would lead to an alternative proof of the classification t h e o r e m to the one provided by [1].

However, since the analysis of the non-closed case requires some additional work, we will present the details elsewhere.

Acknowledgement. T h e author thanks the referee for unusual help in drawing up this corrigendum. T h e author also thanks Professors R. Bryant and G. G r a n t c h a r o v for helpful discussions.

R e f e r e n c e s

[1] BRYANT, R. L., Bochner-K/ihler metrics. J. Amer. Math. Soc., 14 (2001), 623-715.

[2] KAMISHIMA, Y., Uniformization of Ks manifolds with vanishing Bochner tensor. Acta Math., 172 (1994), 299-308.

[3] WEBSTER, S.M., Pseudo-Hermitian structures on a real hypersurface. J. Differential Geom., 13 (1978), 25-41.

YOSHINOBU KAMISHIMA Department of Mathematics Tokyo Metropolitan University 1-1 Minamiohsawa, Hachioji Tokyo 192-0397

J a p a n

k a m i ~ c o m p . m e t r o - u . a c . j p Received February 26, 2004

Received in revised form January 24, 2006

Odkazy

Související dokumenty

In the theory of functions of one complex variable, the proof of the theorem which states t h a t a proper ho]omorphic mapping between domains with real

One importance of finite order maps is t h a t these form the class of transcendental maps which turns up most naturally in the study of the analytic

I n this section we shall show t h a t the deficient values of functions/(z) satisfying the conditions of Theorem 1 give rise to open sets in which the methods

Note t h a t the proof of the representation theorem actually gives some more information than announced in the theorem.. We could however equally well have

Of course, we shall not be able to generalize Nelson's result ([18]) and state t h a t analytic vectors are dense in the space of an arbitrary linear system

Finally, we note the additional corollary below which, in essence, assures us t h a t Theorem 5.1 has non-vacuous content in certain specific circumstances... This

We have noted in the first of the preceding examples t h a t normal (Hermitian) operators are a special class of complex (real) spectral operators on Hilbert space,

Rellich t h a t the most important part of the theory of linear transformations in ~ maintains its validity when t h e condition of separability is dropped,