It is a well-known fact that $S^2 \times S^1$ can be obtained by $0$-surgery on unknot.

What about the $(-1)$-surgery on $S^2 \times S^1$? It seems the resulting manifold, say $W$, bounds contractible manifold.

But I cannot prove it yet or refutes my argument. Any help will be appreciated.


This is true, with some points to clarify. First, you are presumably talking about surgery along a knot that generates the first homology (and hence fundamental group) of $S^1\times S^2$. Then the result of adding the corresponding 2-handle to $S^1\times B^3$ is contractible. The construction (called a `Mazur manifold') goes back to Mazur's paper, A Note on Some Contractible 4-Manifolds, Annals 1961.

The other point is that framing as an integer is not a priori defined for a knot that represents a homology class of infinite order. But fortunately the statement is true for any framing (as in choice of trivialization of the normal bundle). I'd suggest some basic reading about 4-dimensional handle calculus, as in the book of Gompf-Stipsicz.

  • $\begingroup$ Thanks for the reference. To clarify the notion, let me ask you one more question. If we replace $S^2 \times S^1$ by $S^3$ and perform $−1$-surgery on $S^3$ (following your description), may we obtain a manifold bounding rational homology ball? $\endgroup$ – M. Alessandro Ferrari Apr 20 at 10:03
  • 1
    $\begingroup$ Not necessarily; it depends on the knot along which you do that surgery. I actually think these questions are better suited to Mathstackexchange than to this forum. $\endgroup$ – Danny Ruberman Apr 20 at 13:36

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.