Let $r(s)$ be a curve parametrized by arc length, and $\kappa,\kappa',\tau$ are non-zero. Show that $r$ is a spherical curve iff $(1/\kappa)^2+((1/\kappa)'(1/\tau))^2$ is a constant.

The teacher gave the hint "center = $r+(1/\kappa)N+(1/\kappa)'(1/\tau)B$". I know how to get the proof from this, but how can we show that this is the center? Please help. Thanks.

  • $\begingroup$ Show that the distance $\| r(s) - c \|$, where $c$ is the center, is a constant. (Show that its derivative is zero.) $\endgroup$ – Myself Jul 30 '13 at 14:37

Big Hint: The standard technique in all such differential geometry problems is to write $$ r = \lambda T + \mu N + \nu B$$ for some functions $\lambda$, $\mu$, and $\nu$. Differentiate, use what you're given, and use the Frenet equations and use the fact that $T,N,B$ form a basis to get three equations.

  • $\begingroup$ But here I don't have $r$, how can I get $λ,μ,ν$ ? $\endgroup$ – A. Chu Jul 30 '13 at 14:47
  • $\begingroup$ You don't get them explicitly, but you get equations relating them and their derivatives if you follow my suggestion. The crucial thing is: What do you get if you assume the center of the sphere is $0$ and you differentiate $r\cdot r$? $\endgroup$ – Ted Shifrin Jul 30 '13 at 14:53
  • $\begingroup$ I get $λλ'+μμ'+νν'=0$, but I really can't see how I should proceed to the result... $\endgroup$ – A. Chu Aug 1 '13 at 3:45
  • $\begingroup$ No, you're not using definitions of the Frenet frame. What is $r'$? $\endgroup$ – Ted Shifrin Aug 1 '13 at 5:10
  • 2
    $\begingroup$ Awesome! Very powerful technique, and analogues for surface theory too. If you feel like it, accept the answer so the question doesn't appear still to be open. $\endgroup$ – Ted Shifrin Aug 1 '13 at 12:47

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.