CHAPTER V

THE COMPLEX QUADRIC

-1794139851. Outline

This chapter is devoted to the complex quadric which plays a central role in the rigidity problems. In 2-179413585 and 3, we describe the differential geometry of the quadric -179413385 viewed as a complex hypersurface of the complex projective space -179413185. We show that -179413385 is a Hermitian symmetric space and a homogeneous space of the group SO(n + 2). The involutions of the tangent spaces of -179412785 which arise from the second fundamental form of the quadric, allow us to introduce various objects and vector bundles on -179412485 In particular, we decompose the bundle of symmetric 2-forms on -179412285 into irreducible SO(n+2)-invariant sub-bundles; one of these bundles L, which is of rank 2, was first introduced in [18]. In -1794120854, we develop the local formalism of Kahler geometry on the complex quadric following [22]; we wish to point out that auspicious choices lead to remarkably simple formulas. The identification of the quadric -179413385 with the Grassmannian -179411385 of oriented 2-planes in -179411085 given in -1794105855 allows us to relate the geometries of these two manifolds and to define the objects introduced in 3 -179410385in an intrinsic manner. In the next section, we describe the tangent spaces of various families of totally geodesic submanifolds of -179413385 and present results concerning the spaces of tensors of curvature type which vanish when restricted to some of these families. In -1794099857, we determine explicitly the space of infinitesimal Einstein deformations of -179413385 and, from the point of view of harmonic analysis on homogeneous spaces, we compute the multiplicities of a class of isotypic components of the SO(n+2)-module of complex symmetric 2-forms on -179409485 and establish properties of these components. Finally, 8 is devoted to results concerning sections of the sub-bundle -179409285 in 9, we prove that the complex quadric -179413385 is isometric to the product of spheres S2 -179408885× S2.

2. The complex quadric viewed as a symmetric space

-1743746298

-1743746297

-1743746296

for -179407785 If we identify Tb with -179407585 by means of this isomorphism -179407085, since -179406885 for -179406585 the complex structure of Tb is the one determined by the multiplication by i on -179406385 and the Kahler metric g at b is the one obtained from the standard Hermitian scalar product on -179405885 given by (3.6). Moreover, by (5.1) we see that the action of the element

-1743746272

which sends the class -179405285 , where-179405085 , into the point-179404885 is a diffeo-morphism compatible with the actions of G on G/K and X.

The element -179404485 of K belongs to the center of K and is of order 4. The element -179404285 of G determines an involution-179404085 of G which sends G into -179403885. Then K is equal to the identity component of the set of fixed points of -179403685 and (G,K) is a Riemannian symmetric pair. The Cartan decomposition of the Lie algebra -179403385 of G corresponding to -179403185 is

-1743746245

of -179402685, where-179402485 are vectors of -179402285 considered as column vectors. We identify -179401785 with the tangent space of G/K at the coset of the identity element of G and also with the vector space -179401485 in particular, the matrix (5.4) of p0 is identified with the vector -179400785)

If B is the Killing form of g0, the restriction to -179399785 of the scalar product -B is invariant under the adjoint action of K and therefore induces a G-invariant Riemannian metric-179399585 on the homogeneous space G/K. The restriction of Ad j to -179399385 is a K-invariant complex structure on p0 and so gives rise to a G-invariant almost complex structure on G/K. According to Proposition 4.2 in Chapter VIII of [36], this almost complex structure is integrable and the manifold G/K, endowed with the corresponding complex structure and the metric-179399185, is a Hermitian symmetric space. The space G/K is of compact type and of rank 2; when n 3, it is irreducible.

-1743746205

Thus -179398685is a holomorphic isometry from the Hermitian symmetric space G/K, endowed with the metric (1/4n) g0, to X; henceforth, we shall identify these two Kahler manifolds by means of this isometry. According to formula (1.65), it follows that X is an Einstein manifold and that its Ricci tensor Ric is given by

-1743746200

-1793982853. The complex quadric viewed as a complex hypersurface

We begin by recalling some results of Smyth [49] (see also [21]). The second fundamental form C of the complex hypersurface X of -179397885 is a symmetric 2-form with values in the normal bundle of X in -179397885. We denote by S the bundle of unit vectors of this normal bundle.

-1743746190

-1743746188

-1743746187

-1743746186

from the last two equalities, we infer that the relation (5.21) also holds in this case. Thus we have shown that

-1743746183

The remainder of this section is devoted to results of [21] and [23]. We consider the sub-bundle L of -179396485 introduced in [18], whose fiber at -179396285x X is equal to

-1743746176

C according to (5.9), -179395785if Sx, this fiber Lx is generated by the elements -179395585and -179395385and so the sub-bundle L of -179395185 is of rank 2. We denote by -179394985 the orthogonal complement of L in-179394685 By (5.8), we see that L is stable under the endomorphism (1.68) of-179394485 since the automorphism J of T is an isometry, the orthogonal complement-179394285of L in -179394085 is also stable under this endomorphism. We denote by L, L-179393885and -179393685the eigenbundles corresponding to the eigenvalues +i and -i of the endomorphism of -179393385induced by the mapping (1.68). In fact, we have the equalities

-1743746147

-1743746103

-1743746102

-1743746101

for all G and -179388285 moreover-179387785 preserves the submanifold X of -179387585 The group of isometries of X generated by -179387385, which is of order 2, acts freely on X and we may therefore consider the Riemannian manifold -179387085 with the metric gY induced by g, and the natural projection -179386885 , which is a two-fold covering. By (5.33), we see that the action of the group G on X passes to the quotient Y ; in fact, the group G acts transitively on Y . If -179386685 is the subgroup of O(n+2)

-1743746080

where -179386185 it is easily verified that the isotropy group of the point-179385985 is equal to the subgroup -179385785of G. We know that -179385585 is a symmetric space of compact type of rank 2, which we may identify with Y by means of the isometry

-1743746069

sending -179385085 into the point-179384885 for G. Then by (5.33), we see that the projection -179384685 is identified with the natural submersion -179384485of symmetric spaces.

Clearly, we have

-1743746056

-1743746054

-1743746020

4. Local Kähler geometry of the complex quadric

We now introduce the formalism of Kahler geometry on the complex quadric X = -179377085 with n 2, developed in [22, 4].

-1743745984

-1743745983

-1743745982

-1743745981

-1743745980

and -179375485 for 1 j n. Thus since the group G = SO(n + 2) acts transitively on X, if x is a given point of X, from the preceding remark and (5.14) we infer that there exist a section -179375285 of S over a neighborhood -179375085of x and an orthonormal frame -179374785 for the vector bundle -179374585 over -179374385 satisfying -179374185 for 1 j n.

If a is the point -179373785 of S2n+3, then we note that (a) = a; moreover,

-1743745951

-1743745950

-1743745949

-1793731855. The complex quadric and the real Grassmannians

We also consider -179372785 as the standard basis of -179372585. We consider the real Grassmannian -179372085 of oriented 2-planes in -179371885 which is a homogeneous space of G = SO(n + 2), endowed with the Riemannian metric -179371585defined in 1, Chapter IV and denoted there by g; we also consider the homogeneous vector bundles V and W over G-179371385

We define an almost complex structure J on -179370685 as follows. If x-179370485 and -179370285 is a positively oriented orthonormal basis of the oriented 2-plane Vx, the endomorphism J of Vx, determined by

-1743745916

is independent of the choice of basis of Vx and we have J2 = -id. Clearly, the almost complex structure J of -179369485, which is equal to -179369285id on the tangent space (-179369085 of is invariant under the group G. Since -179367885,n4.2 in Chapter VIII of [36], this almost complex structure J is integrable and the manifold -179367385 endowed with the corresponding complex structure and the metric -179367185 is a Hermitian symmetric space.

-1743745885

-1743745883

-1743745821

-1743745779

The sub-bundles -179356085 and (S2T)+- of-179355885 can be defined directly in terms of the intrinsic structure of the real Grassmannian -179355385 without having recourse to the imbedding of X as a complex hypersurface of -179355185 by

-1743745761

From the above inclusions, the relations (5.64) and (5.65), and the decompositions (4.6) and (5.26), we obtain the equalities

-1743745758

We now suppose that n is even. In-179353985 1, Chapter IV, we saw that the oriented 2-plane x-179353785 X determines an orientation of the space Wx. Let x be a point of X and let be an element of Sx. We say that an orthonormal basis -179353585 of -179353385 is positively (resp. negatively) oriented if there is a positively oriented orthonormal basis {v1, v2} of Vx and a positively (resp. negatively) oriented orthonormal basis {w1, . . . ,wn} of Wx such that

-1743745746

for -179352785 Since n is even, it is easily seen that the notions of positively and negatively oriented orthonormal bases of -179352585 are well-defined. Also an arbitrary orthonormal basis of -179352385 is either positively or negatively oriented.

We now consider the case when n = 4. The orientations of the spaces -179351985 a, with -179351785a X, and the scalar product g2 give rise to a Hodge operator

-1743745731

By formulas (3.6) of [21] and (5.68), we easily verify that this auto-morphism of the vector bundle -179351285 is equal to the involution of the vector bundle (-179351085 defined in [21, 3] in terms of an appropriate orientation of the real structures of X. Thus the eigenbundles F+ and F- of this involution of -179350885 corresponding to the eigenvalues +1 and -1, which are considered in [21, -1793506853], are equal to -179350485 respectively. The decomposition +-179350185 and

-1743745715

LEMMA 5.2. Let X be the quadric -179349685 For all -179349485 we have

-1743745708

-1793490856. Totally geodesic surfaces and the infinitesimal orbit of the curvature

We begin by giving an explicit representation of the infinitesimal orbit of the curvature of the complex quadric X = -179348685 with n 3.

-1743745700

for all 2-179348085 From formulas (1.72) and (5.10), it follows that

-1743745691

We now introduce various families of closed connected totally geodesic submanifolds of X. Let x be a point of X and be an element of Sx. If -179347285 is an orthonormal set of vectors of -179347085, according to formula (5.10) we see that the set ExpxF is a closed connected totally geodesic surface of X, whenever F is the subspace of Tx generated by one of following families of vectors:

-1743745681

According to [10], if F is generated by the family (A2) (resp. the family (A3)) of vectors, where {, } is an orthonormal set of vectors of -179346285 the surface ExpxF is isometric to the complex projective line -179345785 with its metric of constant holomorphic curvature 4 (resp. curvature 2). Moreover, if F is generated by the family (A1), where -179345285 is an orthonormal set of vectors of -179344985 the surface ExpxF is isometric to a flat torus.

For 1 j -1793442854, we denote by -179344085 the set of all closed totally geodesic surfaces of X which can be written in the form ExpxF, where F is a subspace of Tx generated by a family of vectors of type (Aj ).

According to 5-179343685, there exists a unit vector v of Vx such that an arbitrary submanifold Z belonging to the family -179343485 can be written in the form Expx-179343285 where -179343085 is a two-dimensional subspace of Wx. We consider the Riemannian metric -179342885 on the Grassmannian -179342585 defined in Chapter IV; by Lemma 4.6 and the relation (5.59), we see that the submani-fold Z is isometric to the Grassmannian -179342385 endowed with the Riemannian metric -179341585 Therefore such a submanifold Z is isometric to a sphere of constant curvature 2 (see also [10]); moreover, by Lemma 4.6 we also see that the image of Z under the mapping -179341385 is a closed totally geodesic surface of Y isometric to the real projective plane endowed with its metric of constant curvature 2.

-1743745627

according to (5.10) we also see that the set ExpxF is a closed connected totally geodesic surface of X. Moreover, according to [10] this surface is isometric to a sphere of constant curvature 2/5. We denote by -179340585F5,the set of all such closed totally geodesic surfaces of X.

If -179340185 is an orthonormal set of vectors of -179339985 and if F is the subspace of Tx generated by the vectors

-1743745613

according to (5.10) we see that the set ExpxF is a closed connected totally geodesic surface of X. Moreover, according to [10] this surface is isometric to the real projective plane -179339085 of constant curvature 1. Clearly such submanifolds of X only occur when n 4. We denote by -179338585 the set of all such closed totally geodesic surfaces of X.

If -179338185 is an orthonormal set of vectors of -179337985 and if F is the subspace of Tx generated by the vectors

-1743745593

according to (5.10) we see that the set ExpxF is a closed connected totally geodesic submanifold of X. Moreover, this submanifold is isometric to the complex projective plane -179337485 of constant holomorphic curvature 4. Clearly such submanifolds of X only occur when n 4. We denote by -179337285 the set of all such closed totally geodesic submanifolds of X.

When n 4, clearly a surface belonging to the family F2,or to the family -179336885 is contained in a closed totally geodesic submanifold of X belonging to the family F7,. In fact, the surfaces of the family F2,(resp. the family -179336485 corresp-179336285ond to complex lines (resp. to linearly imbedded real projective planes) of the submanifolds of X belonging to the family -179336085viewed as complex projective planes.

Let Z be a surface belonging to the family -179335685 with 1 j 6; we may write Z = ExpxF, where F is an appropriate subspace of Tx. Clearly, this space F is contained in a subspace of Tx which can be written in the form -179335185 where F1 is a subspace of -179334985of dimension -179334785; we may suppose that this integer k is given by

-1743745561

According to observations made in 5-179334285, the surface Z = ExpxF is contained in a closed totally geodesic submanifold-179333785 of X isometric to the quadric -179332885 where W1 is a subspace of Wx of dimension k.

Let Fx be the family of all closed connected totally geodesic submani-folds of X passing through x which can be written as-179332185 where

xW1 is a subspace of Wx of dimension 3. We know that a submanifold of X belonging to F is isometric to the quadric -179331185 of dimension 3.

-1743745525

of closed connected totally geodesic submanifolds of X isometric to Q3. We have seen that a surface belonging to the family -179330585 with 1 j -1793303855, is contained in a closed totally geodesic submanifold of X belonging to the family F.

Since-179329985 the group G acts transitively on the set of all maximal flat totally geodesic tori of X and also on a torus belonging to -179329785 and since a surface of -179329585 is a flat 2-torus, we see that, if Z is an element of and if x is a point of Z, there exists an elemen-179329385t Sx and an orthonormal set of vectors -179329185

-1743745505

It follows that the family F1 is equal to .

In [10], Dieng classified-179328485 all closed connected totally geodesic surfaces of X and proved the following:

PROPOSITION 5.3. If n-179328085 3, then the family of all closed connected totally geodesic surfaces of X is equal to -179327885

-1743745466

-1743745465

-1743745463

-1743745462

-1743745442

-1743745430

-1743745429

-1743745424

This inclusion and (5.84) give us the first relation of (5.98). Now suppose that u belongs to -179320585 When n 5, by Lemma 5.7 we know that the relation (5.80) holds; then according to the first formula of (5.94) and Lemma 5.6,(iv), we see that the expression (Tr u-179320085) vanishes. When n = 4, according to the first formula of (5.94) and the equalities (5.87) of Lemma 5.5, we have

-1743745414

vanishes. Hence according to (5.95), the expression (Tr u)-179319485 vanishes. Since Tr u belongs to -179319285 we know that 12528 Thus we have proved assertion (ii).

The following two propositions are direct consequences of Proposition 5.9 and the second equality of (5.74), with j = 2. In fact, Proposition 5.11 is given by Proposition 5.1 of [21].

-1743745402

BB is direct; we also know that it is a sub-bundle of -179318385 Using the relations 2 (1.79) and (5.73), in [18] we were able to determine the ranks of the vector bundles -179317885+ an-179317485 When n 5, by means of Lemmas 5.4–5.7 and other analogous results, in [21] we found explicit bounds for the ranks of the vector bundles -179317285 and -179316685 From these results, the relation (1.79), the second equality of (5.74) and the inclusion (5.75), with j = 2, we obtain the following proposition (see [18,-179316485 5]):

-1743745378

-1793160857. Multiplicities

In this section, we shall suppose that n 3. Let g and k denote the complexifications of the Lie algebras -179315685 and k0 of G = SO(n + 2) and its subgroup K, respectively. Let-179315485and -179315285be the duals of the groups G and K, respectively.

-1743745366

We first suppose that n = 2m, with m 2. We set -179314785 We choose Weyl chambers of -179314585 and (k, t) for which the system of simple roots of g and k are equal to -179314385, and -179314185 respectively. The highest weight of an irreducible G-module (resp. K-module) is a linear form

-1743745355

The equivalence class of such a G-module (resp. K-module) is determined by this weight. In this case, we identify-179313685 (resp. -179313485 with the set of all such linear forms on-179313285

-1743745335

which belong to -179311685 and-179311485 respectively. The complexification -179311285 of -179311085 admits the decomposition

-1743745324

where -179310585 and -179310385 are the eigenspaces of the endomorphism Ad j of p corresponding to the eigenvalues +i and -i, respectively. Since j belongs to the center of K, this decomposition of p is invariant under the action of K on p. We thus obtain the K-invariant decomposition

-1743745317

-1743745293

We consider the subgroup

-1743745267

of G, which we introduced i-179304785n 3 and which contains the subgroup K. The decomposition (5.104) gives us the -179304585 invariant decomposition

-1743745259

-1743745258

-1743745257

-1743745256

According to the equalities (5.108) and (5.109) and Proposition 2.40, since the symmetric space X is irreducible and is not equal to a simple Lie group, we see that E(X) vanishes when n -179303785 4, and that E(X) is isomorphic to the G-module -179303585 when n = 4. By Lemma 5.15 and the Frobenius reciprocity theorem, we see that

-1743745249

if n -179303085 4, then Lemma 5.15 tells us that -179302885 is an irreducible G-module.

-1743745211

When -179299285 the vanishing of the space E(X) is also given by Theorem 1.22 (see Koiso [41] and [42]).

From the branching law for G = SO(n + 2) and K described in Theorems 1.1 and 1.2 of [54], using the computation of the highest weights of the irreducible K-modules given above we obtain the following two propo-sitions:

-1743745203

-1743745202

-1743745158

-1743745157

-1743745156

-1743745110

-1743745109

-1743745107

where b1, b2 are complex numbers which do not both vanish, and that there is a non-zero constant -179288885 such that the relation (5.128) holds. From these remarks, we obtain the following equalities among irreducible G-modules

-1743745102

We no longer assume that n = 4 and return to the situation where n is an arbitrary intege-179288385r 3. Since Hess 12569 is a

-1743745092

-1743745091

-1743745090

-1792872858. Vanishing results for symmetric forms

This section is mainly devoted to results concerning the sections of the vector bundle L and to the proofs of the following two results:

PROPOSITION 5.26. Let X be the complex quadric -179286685 with n-179286485 3. A section h of L over X, which satisfies the relation div h = 0, vanishes identically.

Theorem 5.27. Let X be the complex quadric -179286085 with n-179285885 3. An even section of L over X, which belongs to the space 12571 vanishes identically. Moreover, we have the equality

-1743745069

-1743745068

hold at a. Since the determinant of the matrix

-1743745065

is positive, when r, s -1792846851 the coefficients -179284485 band b vanish, and so in this case h vanishes. Since

-1743745058

when either r = 0 or s = 0, by (5.138) and (5.139) we see that the relation div -179283985 = 0 implies that h vanishes.

-1743745053

-1743745052

-1743745051

-1743745050

-1743744987

satisfies -179276885 From Proposition 5.31, we therefore obtain the relation -179276685 According to our hypothesis, -179276485hbelongs to the space -179276285 and so the other assertion of the proposition is a consequence of Proposition 5.32.

Let -179275585 0 be given integers. According to Proposition 5.33 and the description of the highest weight vectors of the G-module 12583 given above, we see that the space

-1743744961

Since D0 is a homogeneous differential operator, by Proposition 5.18 and the relations (2.1) and (5.133), we see that Theorem 5.27 is a consequence of these results.

-1792741859. The complex quadric of dimension two

We endow the manifold -179269185 with the Kahler metric which is the product of the metrics -179268685 on each factor. It is well-known that the Segre imbedding

-1743744900

We consider the involutive isometry-179268185 of -179267985 defined in -1792674854, Chapter III; according to the commutativity of diagram (3.26), it sends the point-179267285 where -179266985 is a non-zero vector of -179266785 into the point -179266285 where -179266085 is a non-zero vector of -179265885 orthogonal to u. We easily verify that the diagram

-1743744871

is commutative.

Now we consider the diffeomorphism -179264785 defined in 5 and the involutive isometry -179264585 of -179264385 defined in-179263785 1, Chapter IV, which sends an oriented 2-plane of -179263585 into its orthogonal complement endowed with the appropriate orientation. If -179263085 is the involutive isometry equal to the composition 12591 the diagram

-1743744842

The commutativity of the diagram (5.143) is now a consequence of the relations-179262385

-1743744790

..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
3.128.204.5