Consider the following planar curve, given in terms of the coordinates $(X,Z)$. (This is the curve formed by the intersection of the surface $z=\frac{x^2}{a^2}+\frac{y^2}{b^2}$ with the vertical plane $-x\sin\theta+y\cos\theta=0$, and then setting $Z=z$ and $x=X\cos\theta$).


I want to find the curvature of this planar curve at the point $Z=X=0$, and then to find it's maximum and minimum values as $\theta$ varies, where $a\gt b\gt0$.

Here is my attempt so far: we can find a parametrisation for the above curve since it is of the form $Z=Z(X)$, which yields the parametrisation $X\to\begin{pmatrix}X\\Z(X)\\ \end{pmatrix}=:\mathbf c$. Then we can find $\dot{\mathbf c}=\begin{pmatrix}1\\Z'(X)\\ \end{pmatrix}$ and $\ddot{\mathbf c}=\begin{pmatrix}0\\Z''(X)\\ \end{pmatrix}$.

So then, recalling that for such a parametrisation we have $\kappa(X)=\frac{Z''(X)}{[1+(Z'(X))^2]^{3/2}}$, we get that:


So that at $X=Z=0$,


From which it is fairly simple to find the maximum and minimum.

My query: looking over the solutions it seems that I have the right idea, except that there, curvature is given with respect to the parameter $\theta$ as opposed to $X$. What is the significance of choosing $\theta$ over $X$ as the parameter with which to parametrise the given curve? Is there a right or wrong way to choose a parameter when several are involved? Should I have written $\kappa(X,\theta)$ instead of choosing either $X$ or $\theta$? As far as I can see, if I chose to parametrise in terms of $\theta$ over $X$ I wouldn't get the same $c$ it's subsequent derivatives.


As you have stated your problem, $\theta$ is not a valid choice of parameter for the curve obtained by intersecting the surface with the vertical plane $-x \sin(\theta) + y \cos(\theta)$: the curve lies in that plane, $\theta$ is constant on that plane, so $\theta$ is constant on that curve, and a quantity which is constant on a curve cannot be a parameter for that curve. If you change the value of $\theta$, then the vertical plane changes, and its intersection with the surface is an entirely different curve.

But to answer your question, while there are correct parameters for a curve (such as $X$) and incorrect parameters for that curve (such as $\theta)$, there is no single correct choice of parameterization of a curve. For example, if the curve is parameterized as $t \mapsto (x(t),y(y))$ for $a<t<b$, and if $f : (c,b) \to (a,b)$ is any smooth bijection with smooth inverse, then the curve can also be parameterized as $s \mapsto (x(f(s)),y(f(s)))$.

  • $\begingroup$ That makes perfect sense, and I failed to see it like that, thanks. With regards to the expression I have as $\kappa(X)$ am I right in saying that I have calculated this correctly? Or should this be $\kappa(\theta)$? This is what I have in my solutions ($\kappa(\theta))$, but from what you have said, and having moved to find $\kappa(X)$ it doesn't seem it would ever have made sense to find $\kappa(\theta)$. $\endgroup$ Apr 22 '18 at 20:21
  • 1
    $\begingroup$ Rather than go through your calculation, I'll just point out that the expression $\kappa(X)$ does not make sense. Think about it like this. Since for each $\theta$ you have a different curve, and since you are evaluating the curvature of that curve at the point $X=0$, $Z=0$, it follows that $\kappa$ depends on $\theta$, it has no dependence on $X$. So, since $\kappa$ is a function of $\theta$, it should be written $\kappa(\theta)$. $\endgroup$
    – Lee Mosher
    Apr 22 '18 at 21:10

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.