Gabai's solution of the Property R conjecture shows that a minimal genus Seifert surface of a knot, capped off in the 0-framed surgery along that knot, is of minimal genus in its homology class. In particular, it is incompressible in the 0-surgered manifold. On the other hand, there may be incompressible Seifert surfaces for the knot that are not of minimal genus. (For example many pretzel knots bound incompressible surfaces of arbitrarily high genus.) Presumably, there may be such a (non-minimal genus) incompressible surface that becomes compressible in the 0-surgered manifold. Does anyone know an example of this phenomenon?


Revised answer: I found a paper of Ozawa and Tsutsumi which constructs examples of Seifert surfaces which compress when capped off in the 0-framed surgery. See also Tsutsumi for further examples.

In fact, their constructions give knots with Seifert surfaces with "accidental peripheral" curves. That is, given a knot $K$ with exterior $E(K)$, and Seifert surface $S\subset E(K)$, then $S$ has an accidental peripheral curve if it has an essential curve $c\subset S$ (not parallel to the boundary of $S$) which is homotopic into $\partial E(K)$. When one performs 0-framed surgery on $K$ and caps off $S$ to get a closed non-separating surface $\hat{S}\subset S^3_K(0)$, then the curve $c\subset \hat{S}$ is essential in $\hat{S}$ since it is in $S$ and $S$ has only one boundary component, but it will be homotopically trivial in $S^3_K(0)$, since as they point out the homotopy of $c$ to $\partial E(K)$ will be realized by an annulus with interior embedded in $E(K)-S$, and hence intersecting $\partial E(K)$ in a curve of slope $0$ which gets capped off to a disk in $S^3_K(0)$.

Here's an explicit example from Tsutsumi's paper: a knot with accidental Seifert surface

Previous answer:

(this was too long for a comment):

I don't know an answer, but I'll make a suggestion. If you have a knot with infinitely many Seifert surfaces, then there's a closed incompressible surface (it's a kind of limit measured lamination). http://www.ams.org/mathscinet-getitem?mr=2420023

If you perform 0-framed surgery, and the Seifert surfaces remain incompressible when capped off, then this closed surface should also be incompressible in the Dehn filled manifold. So my suggestion is to find a knot with a closed incompressible surface in its complement which compresses in 0-framed surgery. The pretzel knot examples don't seem to work. One possibility is to stick a handlebody with this property in $S^3$, such as http://www.ams.org/mathscinet-getitem?mr=2032111. Then you would also need to check that summing a minimal genus Seifert surface with the closed surface infinitely many times remains incompressible. I'm not sure how to do this part.

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.