9
$\begingroup$

All smooth simple closed curves in $\mathbb R^3$ (knotted or not) can be isometrically embedded into $\mathbb R^2$ as a circle of equal arclength.

The situation for knotted spheres seems more complicated because metric surfaces are more rigid. Here I am unsure that if given a smooth knotted sphere in $\mathbb R^4$ if there is a way to embed into $\mathbb R^3$ while preserving its metric structure. So my question is:

Does there exist a nontrivial smooth knotted sphere in $\mathbb R^4$ which can be smoothly isometrically embedded into $\mathbb R^3$? And if so, are there any concrete examples?

In all of the above, the Riemannian metric structures are assumed to be inherited from the ambient space.

$\endgroup$
8
  • 1
    $\begingroup$ How can a knot be embedded into the plane?? $\endgroup$ Jul 5, 2020 at 5:27
  • 2
    $\begingroup$ There's a slight inconsistency in terminology, but I believe the author at that point in the question is thinking of "circle" as being a synonym for "knot". $\endgroup$ Jul 5, 2020 at 6:01
  • $\begingroup$ Can you clarify the metrics you're using in the isometries, what exactly you mean by a round circle, and be specific about what the analogous situation is when you move up in dimension? $\endgroup$
    – guest
    Jul 5, 2020 at 7:39
  • 1
    $\begingroup$ Thank you for mentioning it can be done C^1 (via the Nash-Kuiper theorem I presume). However, I am mainly interested in maps with more regularity. Smooth would be ideal, but C^2 would be interesting too. $\endgroup$ Jul 7, 2020 at 23:37
  • 1
    $\begingroup$ Maybe you are aware of this, but this paper gives an example of a metric on $S^2$ which does not admit a $C^2$ isometric embedding into $\mathbb R^3$: projecteuclid.org/download/pdf_1/euclid.jdg/1214429999. $\endgroup$ Jul 8, 2020 at 3:51

1 Answer 1

8
$\begingroup$

A spun knot will give a 2-sphere embedded in $\mathbb{R}^4$ whose intrinsic metric embeds into $\mathbb{R}^3$ isometrically as a surface of revolution.

Take a tangle $T$ in $\mathbb{R}^3_+$ with two ends on $\mathbb{R}^2$, that is an embedding $e: ([0,1],\{ 0, 1\}) \to (\mathbb{R}^3_+,\mathbb{R}^2)$. We may assume that $e$ is an isometry. Spin $T$ about $\mathbb{R}^2$ in $\mathbb{R}^4$ to get a surface $S$.

spun knot If $T$ is knotted then $S$ will be knotted. Then there exists an isometric embedding $e’:([0,1],\{0,1\})\to (\mathbb{R}^2_+,\mathbb{R})$ so that $d(e’(t),\mathbb{R})=d(e(t),\mathbb{R}^2)$ for all $t,\in [0,1]$ and thus the surface of revolution $S’$ of $e’([0,1])$ about $\mathbb{R}$ in $\mathbb{R}^3$ is isometric to $S$.

enter image description here

$\endgroup$

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.