What is known about the classification of knots in a solid torus $S^1 \times D^2$? Is enumerating them a reasonable problem? Do we get a similar classification as for knots in $S^3$? Ideally there would be a simple description of the Seifert fibered knots in $S^1 \times D^2$ like for prime non-satellite knots in $S^3$ (they're exactly the torus knots).

The motivation is to better understand the classification of prime knots in $S^3$ as

  1. torus knots,
  2. hyperbolic knots, or
  3. nontrivial satellites,

because satellites come from combining a knot in $S^3$ and a knot in $S^1 \times D^2$.

One issue is that we would want to exclude knots that are only knotted in a ball in $S^1 \times D^2$ because these aren't really new: they come from a knot in $S^3$, and they give composite knots in the satellite construction. Is there a relatively simple way to exclude these?

More formally: one class of knots to exclude are those obtained by taking a $(1,1)$-tangle in a $3$-ball, then closing up the ends so that the knot is not null-homologus in $S^1 \times D^2$. These are nontrivial knots in $S^1 \times D^2$ but they don't really use the topology of the solid torus in an interesting way, so we want to exclude them from our classification.

  • 1
    $\begingroup$ What are Seifert fibered knots? $\endgroup$
    – user101010
    May 21 at 20:11
  • $\begingroup$ Links are classified using the same technology that classify knots. This includes satellite operations, as Sam Nead mentions, these are a special class of 2 (or more) component links. In particular, in the classification it becomes clear that you do not want to think of satellite operations as being generated by knots in solid tori, as that way of generating the operations can't see the symmetries inherent in satellite operations. $\endgroup$ May 22 at 7:08
  • $\begingroup$ @user101010 A Seifert fibered knot is one whose complement is a Seifert fiber space $\endgroup$ May 22 at 20:59
  • $\begingroup$ @RyanBudney Do you have any more details about the "right" way to think about satellite operations that lets you see those symmetries? Maybe a reference? $\endgroup$ May 23 at 12:53
  • 1
    $\begingroup$ Here's my preferred way: arxiv.org/abs/math/0506523 $\endgroup$ May 24 at 3:45

1 Answer 1


Up to Dehn twists, the class of knots in the solid torus is identical to the class of two-component links in the three-sphere, where the first component is an unknot.

For example, Seifert fibered knots in the solid torus give Seifert fibered links in the three-sphere. The base space is the orbifold $S^2(p,\infty,\infty)$.

There is a similar theory of “satellite” links. The knots you wish to exclude (knotted in a three-ball inside the solid torus) are a special case of these. When thought of as links, they are exactly the split links in the three-sphere (where again the first component is the unknot).

Finally, the knot complement in the solid torus is hyperbolic exactly when the link complement in the three-sphere is.


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.