Suppose $M, N$ are compact manifolds with boundary, with $M \subset \text{int}(N)$, and assume that $M$ is diffeomorphic to $N$.

Is $N\setminus M$ automatically diffeomorphic to $\partial N \times [0,1]$? If not, what is a counterexample?

I suspect that the answer to the first question is negative, but I haven't managed to come up with a counterexample despite drawing a lot of pictures, albeit two-dimensional ones. I also attempted to prove that the question has a positive answer, by following the gradient flow of an appropriate smooth function on $N$, but I don't know how to guarantee that such a function has no critical points, and this seems to spoil my proof idea.


One counterexample is an annulus $N=S^1\times [0,1]$ (think of a flat disk with a smaller open disk removed), and $M$ lying inside a small disk in $N$, therefore not boundary-parallel.

enter image description here

  • $\begingroup$ That's an easy one :). Thanks a lot...I was making things too complicated. $\endgroup$ – Matthew Kvalheim Jun 12 '18 at 23:19
  • 1
    $\begingroup$ You're welcome. I'm still wondering about the case when $\partial N$ has to be connected though. $\endgroup$ – Arnaud Mortier Jun 12 '18 at 23:22
  • $\begingroup$ Good point. I'm also curious about that case. $\endgroup$ – Matthew Kvalheim Jun 12 '18 at 23:27

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.