Part 1: The following picture is from Saveliev's book Lectures on Topology of 3-manifolds, page 130:


He indicates that the knot drawn in the solid torus $S^1 \times D^2$ is homologous to $S^1 \times \{ 0\} \subset S^1 \times D^2$.

How can we prove such a claim? Do we know the classification of null-homologous knots in $S^1 \times D^2$?

Part 2: To obtain a Mazur manifold $W$, he says that we attach a $2$-handle to $S^1 \times D^3$ (the dotted circle) with framing $3$. Diagrammatically, we have:

enter image description here

If the knot is not null-homologous, where does the framing $3$ come from?

  • 4
    $\begingroup$ The knot shown is homologous to $S^1 \times \{0\}$ but it is not null-homologous. You might want to fix your question and its title... $\endgroup$
    – Sam Nead
    Commented Mar 15, 2022 at 17:23
  • 2
    $\begingroup$ Why? If not, how can we talk about the framings of given knots? Because the next step is attaching a $2$-handle along this knot. $\endgroup$ Commented Mar 15, 2022 at 17:30
  • 2
    $\begingroup$ I am super confused! I have added a second part to the question... $\endgroup$ Commented Mar 15, 2022 at 17:43
  • 2
    $\begingroup$ You are using the terminology "null homologous" incorrectly. I suggest correcting your question as this will continue to throw people off. Also, why are you asking for a classification of these knots? Seems beside the point of your question. $\endgroup$ Commented Mar 15, 2022 at 18:08
  • 2
    $\begingroup$ Correct, and your curve does not have a Seifert surface in $S^1 \times D^2$, that's why I asked you to try and construct one -- it does not exist. $\endgroup$ Commented Mar 15, 2022 at 20:31

2 Answers 2


Two knots in the solid torus $U = S^1 \times D^2$ are homologous if and only if they have the same (signed) winding number. Proving this boils down to computing the first homology group $H_1(U, \mathbb{Z})$.

For a given oriented knot $K$, its winding number can be computed by (a) finding an oriented meridian disk $D$ for $U$ which is transverse to $K$ and then (b) computing the algebraic intersection number of $K$ and $D$.

Finally you ask: "Do we know the classification of null-homologous knots in $S^1 \times D^2$?" Well, there is a (very complicated) algorithm that decides if two null-homologous knots in $U$ are isotopic. Using this we could produce the (infinite) list of all such knots (said list being complete and without repeats).

where does the framing 3 come from?

I'll guess that he means either (a) the blackboard framing for $K$ or (b) the framing coming from the surface that $K$ and $S^1 \times \{0\}$ cobound. Luckily, these are the same in this particular case.

  • $\begingroup$ For the final part, do you suggest a reference? $\endgroup$ Commented Mar 15, 2022 at 18:18
  • 1
    $\begingroup$ The blackboard framing of a knot $K$ (drawn in the plane) is the normal vector field to $K$, thought of as an immersed curve in the plane. A quick google search found some research articles as well as pictures of framed blackboards, as well as pictures of knots equipped with the blackboard framing. For example: researchgate.net/figure/… $\endgroup$
    – Sam Nead
    Commented Mar 15, 2022 at 18:25
  • 2
    $\begingroup$ The framing comes from the fact that you drew the knot in $S^3$! This is the "magic" idea of Akbulut's carving notation, that allows you to label framings of homologically essential knots $\#^g S^1 \times S^2$ by integers, once you fix a representation of $(\#^g S^1 \times S^2,K)$ in "dotted form" (which is exactly what you have). I have learnt this from Gompf and Stipsicz's book. $\endgroup$ Commented Mar 15, 2022 at 18:33
  • 1
    $\begingroup$ @MarcoGolla Thanks! I have two questions: What does homologically essential knot mean here? Do you suggest section 5.4 of Gompf and Stipsicz's book? $\endgroup$ Commented Mar 15, 2022 at 19:37
  • 1
    $\begingroup$ "Homologically essential" means that its not null-homologous (sometimes with rational coefficients, but in this case it's the same, since there's no torsion). And, yes, I think it's Section 5.4. $\endgroup$ Commented Mar 15, 2022 at 19:55

As @MarcoGolla mentioned, the framing can be controlled due to Akbulut's carving technology: a dotted circle notation. It was introduced in the following article:

Akbulut, Selman. "On 2-dimensional homology classes of 4-manifolds." Mathematical Proceedings of the Cambridge Philosophical Society. Vol. 82. No. 1. Cambridge University Press, 1977.

Let $U$ be the unknot in $S^3$ and $D_U$ be the ribbon disk in $B^4$ with $\partial D_U = U$. Observe that $S^1 \times B^3$ is the ribbon disk exterior of $D_U$, i.e., it is diffeomorphic to $B^4 \setminus \nu(D_U)$ where $\nu(D_U) \approx D_U \times B^2$.

Consider a ribbon knot with a ribbon disk $(K,D) \subset (S^3,B^4)$. Similarly, one can try to understand the $4$-manifold $B^4 \setminus \nu(D)$. The procedure of the construction of the Kirby diagram was answered, for instance here.

Once it is understood, one can also put a dot on the ribbon disk exterior. The reference is again Akbulut's book Section 1.1 and 1.4 about carving ribbon disks:

Akbulut, Selman. 4-manifolds. Vol. 25. Oxford University Press, 2016.

It equivalently represents the ribbon disk exterior. See again Exercise 1.10 and Figure 1.22 in Akbulut's book.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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