Scribd
Upload
28 views5 pages

W Lodzimierz Jelonek and Zbigniew Jelonek

The document discusses a manifold X=S^2×S^3 and shows that it has infinitely many fiber bundle structures over the base B=S^2. Specifically, for each lens space L(p,1) where p is a positive integer, there exists a fibration L(p,1)→X→B. The proof uses properties of lens spaces and algebraic cones to establish the result.

Uploaded by

OBXO
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
28 views5 pages

W Lodzimierz Jelonek and Zbigniew Jelonek

The document discusses a manifold X=S^2×S^3 and shows that it has infinitely many fiber bundle structures over the base B=S^2. Specifically, for each lens space L(p,1) where p is a positive integer, there exists a fibration L(p,1)→X→B. The proof uses properties of lens spaces and algebraic cones to establish the result.

Uploaded by

OBXO
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

MANIFOLD WITH INFINITELY MANY FIBRATIONS

OVER THE SPHERE

WLODZIMIERZ JELONEK AND ZBIGNIEW JELONEK


arXiv:2405.01476v1 [math.DG] 2 May 2024

Abstract. We show that the manifold X = S 2 ×S 3 has infinitely many


structures of a fiber bundle over the base B = S 2 . In fact for every lens
space L(p, 1) there is a fibration L(p, 1) → X → B.

Contents

1. Introduction 1
2. Preliminaries 2
3. Main Result 3
References 4

1. Introduction

Let X, F, B be smooth manifold. Consider smooth fibrations F → X →


B. If X = B = S 1 , then there is infinitely many different fibers Fk such
that we have a fibration Fk → X → B. This suggests that also in higher
dimensions we can find manifolds X, which have infinitely many fibrations
over S 1 . And indeed Tollefson in [10] found some Seifert manifolds having
infinitely many different fiber bundle structures over S 1 . W.P. Thurston in
[9] showed that if a hyperbolic 3-manifold with b1 > 1 fibers over S 1 , then
it fibers in infinitely many ways.
Later Hilden, Lozano, Montesinos-Amilibia in the paper - [7] considered
a certain family of hyperbolic manifolds, obtained as branched covers of
the 3-torus. They showed explicitly that each manifold in this family has
infinitely many different fibrations over S 1 .
Using Tollesfson example one can find diferent examples of this type, if
N is the Tollesfson example and M any manifold then N × M has infinitely
many fibrations over S 1 × M (see [1]). However all examples of this type
have non simply connected base B. To the best knowledge of the authors
there were no examples of a manifold with infinitely many different fibration
over a simply connected base B.
1
2 W.JELONEK AND Z. JELONEK

There is natural question how many different fibration of this type can
exists if the base B is simply connected. Is it possible that the number of
different fibration is infinite? Here using ideas from [4], [5] and [6] we show
the following:

Theorem 3.3. The manifold X = S 2 × S 3 has infinitely many structures


of a fiber bundle over the base B = S 2 . In fact for every lens space L(p, 1)
there is a fibration L(p, 1) → X → B, p = 1, 2, ....

2. Preliminaries

We start with two definitions.

Definition 2.1. Let X ⊂ CPn be an algebraic variety. We assume CPn to be


a hyperplane at infinity of CPn+1 . Then by an algebraic cone C(X) ⊂ CPn+1
with base X we mean the set
[
C(X) = O, x,
x∈X

where O is the center of coordinates in Cn+1 ⊂ CPn+1 , and O, x means the


projective line which goes through O and x. By an affine cone C(X) we mean
C(X) \ X. By the link of C(X) we mean the set L = {x ∈ C(X) : ||x|| = 1}.

Definition 2.2. The three-dimensional lens spaces L(p; q) are quotients of


the sphere S 3 by Z/p-actions. More precisely, let p and q be coprime inte-
gers and consider S 3 as the unit sphere in C2 . Then the Z/p-action on S 3
generated by the homeomorphism

(z1 , z2 ) 7→ (e2πi/p z1 , e2πiq/p z2 )

is free. The resulting quotient space is called the lens space L(p; q).

It is well known that



 Z, k = 0, 3
Hk (L(p, q), Z) = Z/pZ, k=1

0, otherwise

MANIFOLD WITH INFINITELY MANY FIBRATIONS OVER THE SPHERE 3

3. Main Result

Let Wk denotes the Veronese embedding of degree k of CP1 into CPk given
by ψ([z0 , z1 ]) = [z0k : z0k−1 z1 : ...., z0 z1k−1 : z1k ]. Let n ≥ 2 and consider the
varieties Xk = φ(Wk × CP1 ) ⊂ CPnk , where φ : CPk × CP1 → CPnk , φ([z0 :
z1 : ... : zk ], [w0 , w1 ]) = [z0 w0 : z0 w1 : z1 w0 : z1 w1 , ... : zk w0 : zk w1 ] is the
Segre embedding. Consider the affine cone C(Xk ). Let Lk be the link of this
cone. We have

Theorem 3.1. All manifolds Lk are diffeomorphic to S 2 × S 3 = X.

Proof. This result follows from a more general result [4], however for the
convenience of the reader we give here a direct proof of this fact (which
is slightly simpler than this in [4]). By construction, Xk is the union of
projective lines Xk = a∈Wk φ({a} × P1 ). This means that the projective
S

cone C(Xk ) is the union of planes which have the line φ({a} × P1 ) at infinity
and go through the point O = (0, ..., 0). Thus the link Lk, of this cone is a
union of 3-spheres S 3 . In fact using the Ehresmann Theorem, it is easy to
∼ S 2 with the projection
observe that this link is a sphere bundles over Wk =
being the composition of the projection p : CPnk +1 \ {0} → CPnk and the
projection q : Wk × CP1 → Wk . From the homotopy sequence

0 = π2 (S 3 ) → π2 (Lk ) → π2 (S 2 ) = Z → π1 (S 3 ) = 0 → π1 (Lk ) → π1 (S 2 ) = 0,

we get π1 (Lk ) = 0 and by the Hurewicz theorem H2 (Lk , Z) = π2 (Lk ) = Z.


Note that H2 (S 2 ×S 3 ) = Z. Since w1 (S 2 ), w1 (S 3 ) and w2 (S 2 ), w2 (S 3 ) vanish
by the Whitney Product Theorem the Stiefel-Whitney class w2 (S 2 × S 3 ) =
Σ2i=0 wi (S 2 ) × w2−i (S 3 ) = 0. Hence by the Smale-Barden classification (see
[2]) it is enough to prove that w2 (Lk ) = 0.
Let π : Lk → CP1 ×CP1 = B be the Hopf fibration. Let E = ker dπ. Then
E is a one dimensional (real) vector subbundle of T Lk . Let us introduce
on Lk a Riemannian metric g and let F = E ⊥ . Then dπ : F → T B
induces the isomorphism of fibers: dπ : Fx → T Bπ(x) . We will show that
F is isomorphic to π ∗ T B. For a X ∈ Tπ(x) B let Xx∗ ∈ Fx ⊂ Tx Lk be
a vector such that dπ(X ∗ ) = X. The isomorphism is Φ : π ∗ T B → F
given on a fiber π ∗ T Bx as Φ(X) = Xx∗ . Hence T Lk = E ⊕ π ∗ (T B) and
w2 (Lk ) = w2 (F ) + w1 (F ) ∪ π ∗ w1 (B) + π ∗ w2 (B).
Since w1 (S 2 ) and w2 (S 2 ) vanish, by the Whitney Product Theorem, it
follows that w1 (B) = Σ1i=0 wi (S 2 ) × w1−i (S 2 ) = 0, w2 (B) = w2 (S 2 × S 2 ) =
Σ2i=0 wi (S 2 ) × w2−i (S 2 ) = 0, which implies w2 (Lk ) = 0. 
4 W.JELONEK AND Z. JELONEK

Theorem 3.2. Let W ⊂ CPn be a smooth rational curve of degree d. Let


C(W ) be an affine cone with base W. Then the link L of this cone at 0 is
diffeomorphic to the lens space L(d, 1).

Proof. Note that L is a principal circle bundle over the sphere. By [3] (see
Theorem 2.1, Theorem 2.2 and Theorem 2.3) the group of such bundles
is cyclic and generated by S 3 = L(1, 1). Moreover, pS 3 = S 3 /Gp , where
Gp = Z/(p). Hence L is diffeomorphic to some pS 3 , which is the lens space
L(p, 1). By [5], Thm. 3.5 we have H 2 (L, Z) = Z/d. On the other hand
by [8], point 11, the space L is a rational homology 3-sphere, in particular
H1 (L, Z) is a torsion group. Since the torsion part of H1 (L, Z) coincides
with the torsion part of H 2 (L, Z) we have H1 (L, Z) = Z/(d). Hence p = d
and L = L(d, 1). 

Theorem 3.3. The manifold X = S 2 × S 3 has infinitely many structures


of a fiber bundle over the base B = S 2 . In fact for every lens space L(p, 1)
there is a fibration L(p, 1) → X → B, p = 1, 2, ....

Proof. Let Wk denote the Veronese embedding of degree k of CP1 into CPk .
Let n ≥ 2 and consider the varieties Xk = φ(Wk × CP1 ) ⊂ CPnk , where φ
is the Segre embedding. Consider the affine cone C(Xk ). Let Lk be the link
of this cone. By construction, Xk is the union of projective rational curves
S
Xk = a∈CP1 Wk × {a}. This means that the affine cone C(Xk ) is the union
S
of cones a∈CP1 C(Wk × {a}). Thus by Theorem 3.2 the link Lk of this cone
is a union of lens spaces L(k, 1). In fact using the Ehresmann Theorem, it is
easy to observe that this link is a bundle over CP1 ∼ = S 2 with the projection
being the composition of the projection p : CPnk +1 \ {0} → CPnk and the
projection q : Wk × CP1 → CP1 . Consequently Lk is a L(k, 1) bundle over
S 2 . By Theorem 3.1 this finishes the proof. 

References
[1] Alcaraz, J., Ph.D-dissertation, UNIVERSITY OF CALIFORNIA RIVERSIDE, 2022.
[2] Barden, D.; Simply connected five-manifold, Ann. of Math. 82 (1965), no. 2, 365-385.
[3] Blair, D.E., Riemannian Geometry of Contact and Symplectic Manifolds, Second
Edition, Progress in Mathematics Volume 203, Birkhäuser (2010).
[4] Birbrair, L.; Fernandes, A.; Sampaio, J. E. and Verbitsky, M. Multiplicity of singu-
larities is not a bi-Lipschitz invariant. Math. Ann., vol. 377 (2020), 115-121.
[5] Bobadilla, J.F. de; Fernandes, A. and Sampaio, J. E. Multiplicity and degree as bi-
lipschitz invariants for complex sets. Journal of Topology, vol. 11 (2018), 958-966.
MANIFOLD WITH INFINITELY MANY FIBRATIONS OVER THE SPHERE 5

[6] Fernandes, A, Jelonek Z., Sampaio J. E., Bi-Lipschitz equivalent cones with different
degrees, arXiv:2309.07078.
[7] Hilden, H.M., Lozano, M.T., Montesinos-Amilibia, J.M., On hyperbolic 3-manifolds
with an infinite number of fibrations over S 1 . Math. Proc. Cambridge Philos. Soc.
140 (2006), no. 1, 79-93
[8] Kollar, J., Links of complex analytic singularities. In: Surveys in Differential Geom-
etry XVIII. International Press of Boston, (2013), 157–192.
[9] Thurston. W, A norm for the homology of 3-manifolds, Memoirs of the American
Mathematical Society, 59(339):99–130, 1986.
[10] Tollefson. J., 3-manifolds fibering over S 1 with nonunique connected fiber, Proc. Amer.
Math.Soc. 21 (1969), 79–80.

(Wlodzimierz Jelonek) Katedra Matematyki Stosowanej, Politechnika Krakowska,


Warszawska 24, 31-155 Kraków, Poland.
E-mail: [email protected]

(Zbigniew Jelonek) Instytut Matematyczny, Polska Akademia Nauk, Śniadeckich


8, 00-656 Warszawa, Poland.
E-mail: [email protected]

You might also like