Let $K$ be a knot in a solid torus. Combining results of Berge and Gabai, we know that if $K$ admits a solid torus surgery, then $K$ is a 1-bridge braid. Using Gabai's result, we can figure out what the surgery slope is. If a 1-bridge braid admits a non-trivial solid torus surgery, the surgery slope is (almost) always unique (there is a unique 1-bridge braid which admits 2 non-trivial solid torus surgeries; most refer to it as K(7,2,4) in (w,b,t) notation).

If $B = K(7,2,4)$, we know that $B$ is equivalent to its dual, $B'$.

What is known about the relation between a 1-bridge braid and its dual in general? Namely:

Question #1: Given a 1-bridge braid $K$ that admits a solid torus surgery, when is it equivalent to its dual?

Question #2: Is it known when, given 1-bridge braids $K_1$ and $K_2$ such that $K_1 \not \simeq K_2$, their exteriors (in the solid torus $\mathbb{D}^2 \times S^1$) are homeomorphic?

I know Berge gives a necessary & sufficient condition for equivalence of knots, but I'm wondering if the answers to my above questions already exist in the literature in a more modern/non-G-pair-theoretic language.



1 Answer 1


For Question #1:

As you acknowledge, Berge already gives an answer to Question #1. Using his notation, this is summarized in Table 2 of his paper "The knots in $D^2 \times S^1$ which have nontrivial surgeries that yield $D^2 \times S^1$". Berge also already tells you how to translate from a G-pair to a standard braid form in his Lemma 2.2. So really, you just need to do the legwork to get it into a more common presentations of the dual.

Alternatively, a non-G-pair exhibition of the pairs of dual Berge knots is given in the setting of their tangle quotients in Baker-Buck, "The classification of rational subtangle replacements between rational tangles". Figures 3-6 show bandings (+isotopy) between tangles in plat form keeping track of the dual arc, starting from a tangle isotopic to the crossingless two cup tangle.

(In [Baker-Buck], since they’re dual, I had combined types III and V into one family, though I really shouldn’t have. With one exception, there’s no overlap of them (as noted in Table 2 of [Berge]). Referring back to my "Surgery descriptions and volumes of Berge knots II" one finds that in Figure 5 of [Baker-Buck]: Top row shows V to III, Bottom row shows III to V.)

For Question #2:

If the exteriors of $K_1$ and $K_2$ are both homeomorphic to the manifold $X$, then either

In Theorem 3.6 of [Wu], Wu determines both the 1-bridge braids for which a filling of the outer torus boundary of the exterior produces a solid torus again and the slopes of these fillings. What’s still missing is the determination of the dual knot the resulting solid torus: Is it necessarily a 1-bridge braid? If so, which one?

Note that if you drop the assumption of 1-bridge braid, situation (B) becomes much more varied. You’re basically looking at two component links in lens spaces where each component is isotopic (individually) to a core of a Heegaard solid torus. For example in $S^3$ any two component link of unknots has this property.

  • $\begingroup$ This is very helpful, thank you so much! $\endgroup$
    – Krishna
    Jun 6, 2017 at 21:08

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.