I am curious if the area of an ellipse can be shown to be proportional to the integral of all cross-ellipse distances. Before I define cross-ellipse distance, I will give a motivating example from a circle.

Suppose you have a circle with diameter $d$. Then $A_\text{circle}=\pi d^2/4$, and $C=\pi d$ . Now consider a line tangent to the edge of the circle that touches a point $p$ on the edge. The maximal distance from $p$ to any other point on the edge of the circle is equal to $d$. For a given $p$, call such a distance $e(p)$. If I integrate over all these distances, $e_\text{circle}(p)$, I get (perhaps erroneously) \begin{equation} \int e_\text{circle}(p) \, dp = Cd = \pi d^2 = 4A_\text{circle} \end{equation} Here, this seems simple since the maximal distance of any $p$ to another point on the hull of the circle is always $d$.

I am curious about whether something regarding the proportionality of this sort of distance to the area can be said for an ellipse.

Given a point $p$ on the hull of an ellipse, the maximum distance between $p$ and any other point on the hull is to the point $p_\text{far}$ such that a line passing through $p$ and $p_v{far}$ is orthogonal to the tangent touching $p$. Denote such a distance $e_\text{ellipse}(p)$. On a circle $e(p)=d$ but here it varies with $p$.

The area of an ellipse is $A_\text{ellipse} = \pi a b$, where $a$ and $b$ are the semi-major and semi-minor axes ($\frac 1 2$ of the ellipse's major and minor axes), respectively.

Is \begin{equation} \int e_\text{ellipse}(p) \, dp \propto A_\text{ellipse} \end{equation}

Can I show this by using linear combinations of $a$ and $b$ to characterize the distances $e_\text{ellipse}(p)$?

If it is true that the integral of the $e(p)$ is proportional to the area for any convex shape, I would really appreciate an argument or source.

In case it helps to give context, I am interested in this approach as a way to show that for two ellipses with $a_1, b_1$ for the first and $a_2, b_2$ for the second that using this approach can show that the area of one of the ellipses is larger than the area of the other ellipse.

  • $\begingroup$ Correction: line passing through $p$ and $p_{far}$ is orthogonal to the tangent touching $p_{far}$, not $p$. $\endgroup$ – user147263 Jul 4 '14 at 1:16

No, this is not true. Take the ellipse with semiaxes $a$ and $a^{-1}$ where $a$ is very large. It looks like a long thin needle. The area is $\pi$, independent of $a$. But the integral $\int e(p)\,dp$ tends to $\infty$ as $a\to \infty$. Indeed, $e(p)\ge a$ for all $p$ (consider the distances from $0$ to the pointy vertices of ellipse), and the perimeter also grows, it's at least $4a$.


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.