3. Some examples [04SF]
Original official author HTML, exact retained edition. Historical TeX conversion verdicts remain unchanged. Cited-edition alignment and mathematical self-containment are not assessed.
Complete original source context · Original author HTML
3. Some examples
3.1. Riemann surfaces
Let be a closed Riemann surface of genus . It is well-known that admits a decomposition into pairs-of-pants. Namely, there exist disjoint embedded circles such that is a disjoint union of copies of the pair-of-pants . The pair-of-pants surface is homeomorphic to the Riemann sphere punctured in three points.
To such a decomposition we associate a graph . The vertices of correspond to the pairs-of-pants while the edges correspond to the circles . Each edge joins the vertices corresponding to the adjacent pairs-of-pants.
There exists a fibration such that the circles are inverse images of the midpoints of the edges of . Such fibration is canonically associated to our decomposition into pairs-of-pants. To construct it we fiber each individual pair-of-pants over a tripod graph as pictured on the left-hand-side of Figure 5.

3.2. The elliptic curve and the K3-surface
Here we consider the well-known fibrations of the elliptic curve and the K3-surface.
Let be an elliptic curve, i.e. a Riemann surface of genus 1. Since is topologically a torus, there is a trivial -fibration .
Suppose that the elliptic curve is presented as a curve in a toric surface , where is the Newton polygon of a polynomial defining . By the genus formula (see [8]), contains a unique lattice point. By Proposition 1.10 a dual -complex is homotopy equivalent to a circle. It is easy to see that the fibration from Theorem 1’ coincides up to homotopy with the trivial -fibration .
Another famous fibration has the K3-surface as its total space. All its fibers, except for 24 of them are Lagrangian tori.
Suppose that the polygon has exactly one interior lattice point. Then, by Khovanskii’s formula [8], the zero locus of a generic polynomial with the Newton polygon is a K3-surface. A dual -complex is homotopy equivalent to a sphere by Proposition 1.10.
Again, the fibration can be deformed to a fibration like by so-called shelling of 11 1 A higher-dimensional version of such deformation will be the subject for a future paper..
In higher dimensions, if is a non-singular polyhedron with a unique interior lattice point, then the corresponding hypersurface is a smooth Calabi-Yau manifold. Singular torus fibrations were constructed by Zharkov [17]. Ruan [13] noted that such fibrations can be made Lagrangian.
Theorem 1’ constructs in this case a stratified torus fibration over a polyhedral complex homotopy equivalent to .
3.3. Hyperplanes in the projective space
This is a fundamental example for the main theorems. Let be a hyperplane. Its toric part is an open pair-of-pants.
Let be the moment map for (see (3)).
Lemma 3.1.
.
Proof.
By [12] is a spine of the amoeba and, therefore, its subset. The lemma can alternatively be verified by writing explicit inequalities defining . ∎
The complement consists of components. Each component is the region where one of the functions is maximal. In the component corresponding to we consider the foliation into straight lines parallel to the gradient of (the th basis vector). In the component corresponding to we consider the foliation into straight lines parallel to . These foliations glue to a singular foliation which has singularities at .

It is easy to smooth out (in a symmetric way with respect to the homogeneous coordinates permutations) at the open -cells of (see Figure 6). However, the singularities at the smaller-dimensional cells are essential. The leaves passing through an open -cell are homeomorphic to the cone over points.
We denote the resulting foliation with . The foliation is a singular fibration and defines the projection .
The following statement is a key lemma in the proof of the main theorems of this paper.
Lemma 3.2.
The proof of this lemma occupies the rest of this subsection.
To figure out the fibers of we need to understand the critical points of . Following [6] and [11] for a hypersurface we define the logarithmic Gauss map
by taking the composition of a branch of a holomorphic logarithm of each coordinate with the conventional Gauss map. This produces the following formula
where is the polynomial defining .
Note that the Newton polyhedron of coincides with the Newton polyhedron of . Therefore, by Kouchnirenko’s formula [9], . In particular, if then .
Lemma 3.3 (cf. Lemma 3 of [11]).
The set of critical points of coincides with .
Proof.
Let and let be a branch of a holomorphic logarithm defined in a neighborhood of . The point is critical for iff and the orbit of the real torus are not transversal at . But takes the tangent space to an orbit of to a translate of in .
Therefore, is critical iff contains at least purely imaginary vectors which is, in turn, equivalent to . ∎
Corollary 3.4.
The set of critical points of coincides with the real locus of (i.e. with the set of real solutions of ).
Proof.
Note that, since is defined over , we have . Note that extends to a map which is an isomorphism, since . ∎
Corollary 3.5.
The locus of critical values of is an immersed manifold transverse to the foliation .
Proof.
The map is an immersion since the map is an immersion (it is a trivial -covering of ).
To see the transversality we recall the definition of the foliation . For each component of the foliation is parallel to a vector normal to a facet of the Newton polyhedron of . Therefore, any hyperplane in the image is transverse to . Furthermore, hyperplanes close to being parallel to are close to the hyperplane in corresponding to this facet and therefore are far from the given component of . Thus the result of smoothing is also transverse to and the angle between them in is separated from 0. ∎
Note that is a stratified -fibration. Thus, the transversality of and implies that is a stratified fibration for . We need to show that the restriction of to open -cells of is a torus fibration.
Consider a point for a large . Note that is almost horizontal near . Thus the fiber of over is diffeomorphic to the fiber of a composition of and the linear projection onto the first coordinates. Note that the map obtained by taking the arguments of the first coordinates is a diffeomorphism. Recall that is given by the equation . The absolute values of the coordinates are fixed. For any value of their argument we take to get the unique point from corresponding to this choice of the arguments. Since are small .
We verify the conclusions of Theorem 3 item-by-item. The first and the last conclusions are vacuous in this case, since (and, therefore, as well) is contractible. The second one holds since is itself an open pair-of-pants.
To make the third conclusion true we have to modify a little. The fiber is not Lagrangian, but it is close to a Lagrangian torus . We can deform a little in a neighborhood of to make it intersect the fiber of along . Therefore, is Lagrangian for a nearby symplectic structure. By Moser’s trick (see e.g. [2]) there exists a self-diffeomorphism of constant outside of a neighborhood of and taking one symplectic structure to another. We redefine as . This ensures a Lagrangian fiber over one of the open -cells of . We do the same for all other -cells.
3.4. A localization of the standard hyperplane
The toric part of a hyperplane from 3.3 is a nice embedding of to . However for our purposes it is convenient to modify it in a neighborhood of infinity to get a different submanifold which is better suited for gluing.
Note that the symmetric group acts on by interchanging the functions . This action is inherited from the action of on interchanging the homogeneous coordinates since is invariant. The hyperplane is invariant with respect to this action.
Proposition 3.6.
There exists a proper submanifold such that
- •
is embedded to symplectically, i.e. so that the restriction of the form (2) to is a symplectic form.
- •
is isotopic to in .
- •
The composition is a stratified -fibration that satisfies to all hypotheses of Theorem 3.
- •
the closure of in is a smooth manifold isotopic to .
- •
is invariant with respect to the action of the symmetric group on (see above).
- •
For a sufficiently large
where and . In particular, the intersection is invariant under a translation , .

Proof.
We construct inductively by dimension . If then is a point and . Assume that , is already constructed. Consider the simplex
Each its -dimensional face is dual to a -cell of . Fix a sufficiently large number .
First we define . Each -face of is contained in a unique affine -space in . Furthermore, the adjoint faces cut the polyhedron . Thus we may identify with and, therefore, with . By the induction assumption we already have . We define to be equal to the union of these for all faces of . By the induction hypothesis (and since was large enough) the choices over different faces agree.
Our next step is to extend to the complement of . For each face of consider its outer normal cone (e.g. if is a facet then is a ray). We define
In other words, we span the region above the normal cone of a -face by the translates of the manifold .
We set . By now we have defined everywhere, but .
Consider a facet of , e.g. the one sitting in the hyperplane . Since is large enough, is small enough and the intersection is close enough to the zero set of . By the induction hypothesis this zero set can be deformed to . We define , using this deformation. We repeat the same procedure for all other facets of . ∎
Denote . This is “the kernel” of and is diffeomorphic to a closed pair-of-pants (as a manifold with boundary and corners).