[Update: I’ve written quite a bit more about this theorem since 2009. See this page for more details.]
I’ve been playing with GeoGebra for the last few days. As an exercise I decided to create applets to demonstrate the extremely beautiful Japanese Theorem.
The first appearance of the Japanese theorem was as a Sangaku problem. Sangaku problems were geometric problems posted in Buddhist temples and Shinto shrines in Japan as offerings to the gods. They were created during the Edo period when the nation was closed to the outside world.
Essentially the problem amounted to the following. Inscribe a quadrilateral in a circle. Draw one of the diagonals and inscribe a red circle in each of the two resulting triangular regions. Then repeat with the other diagonal, creating two blue circles (see image below). The amazing fact is that the sum of the radii of the red circles is the same as the sum of the radii of the blue circles. Here’s a GeoGebra applet illustrating this behavior. (Incidentally, there is another cool fact about this construction: the centers of the four circles form a rectangle!)
Once the quadrilateral version of the Japanese theorem has been established it is not difficult to extend it to general cyclic polygons. Take any cyclic polygon and triangulate it using nonintersecting diagonals. Inscribe circles in each of the triangles. Then the sum of the radii is independent of the choice of triangulation. For example, the sum of the radii of the blue circles below is equal to the sum of the radii of the red circles. Here’s a GeoGebra applet illustrating this behavior.
Read more about the Japanese Theorem here, here, and here. For more about Sangaku problems see Hidetoshi and Rothman’s new book Sacred Mathematics: Japanese Temple Geometry.
David Richeson: Division by Zero 
That is neat. I think I will add proving it to my summer todo list. Thank you.
I love it — thank you — from a math tutor always interested in new (to me) constructions
It is a cool problem. Be warned, Kate, that in Hidetoshi and Rothman’s book they have chapters titled Easier Temple Geometry Problems, Harder Temple Geometry Problems, and Still Harder Temple Geometry Problems. This problem is in that third chapter.
It is tricky to prove from basic geometric results, but if you happen to know this geometric theorem http://cli.gs/UBZX06 (I’ve shortened the URL in case you didn’t want the proof “spoiled”), it has a quick proof.
Very cool! How do you insert text into GeoGebra that includes variables from the construction – like your “Sum of radii …” text?
You can find instructions for that here under “mixed text.”
Thank you!
WOw.. it’s great… I have studied Circumscribable Quadrilateral just last summer and now I meet inscribable quadrilateral with a great characteristic.. Has this been proven?i’ll try to make some theorems and proofs for this..:)
Check out this other more recent page.