7
$\begingroup$

Recall that one way of drawing closed 2-manifolds is to take a disk $D^2$, take a cellular decomposition of $\partial D^2$, pair the vertices in this cellular decomposition so that the pairing preserves edges, and then take $D$ together with this quotient of the boundary.

We can do this in other dimensions as well, for example in dimension 3, every closed 3-manifold can be obtained by a similar procedure where we take $B^3$, take a cellular decomposition of $\partial B^3$, pair the vertices of this cellular decomposition so that the pairing preserves edges and faces, and then look at the quotient of $B^3$ by this pairing.

Threlfall and Seifert did this for the Poincaré homology sphere (see for example here - which also contains a different such description due to Kreines). In fact, they take the cellulation of $\partial B^3$ to be the dodecahedron. Is there a complete (presumably rather short) list of all of the 3-manifolds obtained in such a way where the cellulation is a Platonic solid? $T^3$, $\mathbb{R}P^3$, and the Seifert-Weber space are other examples that come to mind. I'd guess that the Poincaré homology sphere is maybe the only homology sphere on that list. More generally, I'd like to look through a list of the 3-manifolds that occur in this way using simple cellulations.

This can also be done in a similar way in dimension 4 to produce all smooth closed 4-manifolds. Are there some nice pictures/examples of this being carried out somewhere? I'd love to see such pictures of $S^2 \times S^2, T^4, \mathbb{C}P^2,...$

$\endgroup$
3
  • $\begingroup$ In 3d you can find some references here. In 4d, I think, the most common way to describe 4-manifolds (at least, among topologists) is via Kirby diagrams or trisections. $\endgroup$ Commented Dec 9, 2020 at 18:32
  • $\begingroup$ In 3d see arxiv.org/abs/0806.1912, "Bitwist 3-manifolds" by J. W. Cannon, W. J. Floyd, W. R. Parry, which constructs all closed connected orientable 3-manifolds by identifying sides of a 3-disk. $\endgroup$ Commented Dec 10, 2020 at 14:29
  • $\begingroup$ @IgorBelegradek The way I think of doing this for arbitrary such 3-manifolds is to start with a triangulation and consider the graph whose vertices are 3-cells and the edges correspond to the faces between adjacent 3-cells. Taking a spanning tree of this grpha and the taking all of the 3-cells and 2-cells corresponding to the vertices and edges gives a 3-ball with the desired quotienting of the boundary. $\endgroup$
    – user101010
    Commented Dec 10, 2020 at 20:21

1 Answer 1

5
$\begingroup$

This closed orientable 3-manifolds obtained by gluing faces of the Platonic solids were classified by Everitt.

That is for regular polyhedra with equal dihedral angles, and the gluing is done geometrically. However, it is also possible to do the gluing topologically, and for that problem, I only have a partial answer. There are 3 closed orientable 3-manifolds obtained by gluing faces of the tetrahedron. They are $S^3$, $L(4,1)$, and $L(5,2)$. Explicit gluings can be seen in figure 2 of this paper of Jaco and Rubinstein.

There are 17 closed orientable 3-manifolds obtained by gluing faces of the octahedron, 13 of which are prime. They are listed in Proposition 4.2 of this paper by Heard, Pervova, and Petronio.

Presumably, the closed orientable 3-manifolds obtained from the cube have been enumerated, but I don't know of a reference. They include $\mathbb{R}P^3$, the 3-torus and at least 2 of the other closed orientable Euclidean 3-manifolds. I imagine there are a lot of 3-manifolds obtained from the dodecahedron and icosahedron, but I doubt that anyone has enumerated them all.

As for 4-manifolds, I will leave that for someone else to answer, except to note that there are no 4-manifolds obtained from a single pentachoron (4-simplex), since it has 5 tetrahedra in its boundary and this causes a parity issue.

$\endgroup$
1
  • $\begingroup$ Thanks this was very helpful! A quick look through those references does seem to suggest that the Poincaré homology sphere is so far the only homology sphere (other than $S^3$) that is popping up. $\endgroup$
    – user101010
    Commented Dec 10, 2020 at 9:30

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .