If we attach a $4$-dimensional $1$-handle $D^1 \times D^3$ to a $4$-dimensional $0$-handle $B^4$, we obtain $S^1 \times D^3$. The null homologous knot in $S^1 \times S^2$ indicated in the picture lives in a solid torus/attaching region $S^1 \times D^2$ and $$ S^1 \times D^2 \hookrightarrow S^1 \times S^2 \hookrightarrow S^1 \times D^3.$$

In pg. 356 of his book Knots and Links, Dale Rolfsen notes that

  1. $\pi_1(W) = 1$,
  2. $\pi_1(\partial W) = \langle x,y \ \vert \ y^5 = x^7, y^4 = x^2 y x^2 \rangle \neq 1$.

How we compute the relevant fundamental groups using the following diagram? What is the strategy?

enter image description here

  • 4
    $\begingroup$ Handle attachments are basically "structured" cw-complexes. There is a standard process to compute the fundamental group of CW-complexes described in algebraic topology textbooks. For the boundary there are a few ways. One would be to start with the Wirthinger presentation of the link exterior, then attach the appropriate relators as one fills. I think Rolfsen explains this, a little earlier in his book. $\endgroup$ Feb 19 at 0:19
  • 2
    $\begingroup$ I don't understand. We can apply the Wirtinger's method to the knots in $S^3$. In OP, knots live in $S^1 \times S^2$. Is there a way generalize this process to another $3$-manifolds? $\endgroup$ Feb 19 at 10:46
  • 1
    $\begingroup$ Yes, $S^1 \times S^2$ you can obtain as $0$-surgury on the unknot, so your surgury diagram for $\partial W$ will consist of a union of the surgery curves in your diagram with a $0$-labelled unknot. Going around the $S^1$ factor of $S^1 \times S^2$ corresponds to linking the unknot. $\endgroup$ Feb 19 at 19:57
  • $\begingroup$ Great! To be hundred percent sure, would you sketch a diagram, please? $\endgroup$ Feb 19 at 23:01
  • $\begingroup$ Plus, do we generalize your technique to any surgered $3$-manifold, or $S^1 \times S^2$ is special here? Do you have a reference (providing proofs) for generalizing process of Wirtinger presentations? $\endgroup$ Feb 19 at 23:03

1 Answer 1


Thanks to the dotted circle notation of Professor Akbulut, we can easily (compare with your picture) draw the handle diagram of a typical Mazur manifold as follows:


Since $1$-handles and $2$-handles respectively give the generators and relations of the fundamental group of a $4$-manifold, we have (for my $W$): $$\pi_1(W) = \langle \alpha \ \vert \ \alpha^2 \alpha^{-1} \rangle = 1.$$

The boundary $3$-manifold $\partial W$ is already in the picture. Further, we can compute $\pi_1(\partial W)$ from the diagram by using Wirtinger's presentation.

Because today is Sunday (which makes me so lazy) and Eylem already did such a computation (another purpose but for a Mazur manifold) in Example 1 in [Yıldız, 2018], the rest is for you, mimic the argument by keeping Ryan Budney's comments in your mind:


Bonus: In particular, $W$ is a contractible $4$-manifold: By using Mayer-Vietoris sequences, observe that $W$ is a homology $4$-ball, i.e., we have $H_*(W, \mathbb{Z}) = H_*(B^4, \mathbb{Z})$. Then apply the classical theorems of Hurewicz and Whitehead.

Bonus 2: Since $W$ is contractible, $\partial W$ must be a homology $3$-sphere, i.e., $H_*(\partial W, \mathbb{Z}) = H_*(S^3, \mathbb{Z})$. Using Kirby calculus, show that $\partial W \approx \Sigma(2,5,7)$ where $\Sigma(p,q,r)$ denotes the Brieskorn sphere given coprime positive integers $p,q$ and $r$.


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.