Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

How can one go about solving the following problem?

Inscribe a circle in an arbitrary triangle. Call it's radius $r_1$. Inscribe three more circles so that each one is tangent to two sides of the triangle and the first circle (i.e., each at a different corner). Call the radii $r_2, r_3, r_4$. Find a relationship between $r_1, r_2, r_3$ and $r_4$.

The most promising method of attack for me was to consider the isosceles triangles at each corner: the base being the tangent line to the point of intersection of the angle bisector of the triangle and the first circle. But I'm stuck.

Any suggestions much appreciated.

share|cite|improve this question
Have you considered the simpler case of inscribing a circle in an equilateral triangle of side-length $1$? In this case we will have $r_2 = r_3 = r_4$. I am not promising a general formula will pop out here, but when I am stuck on a problem, I always consider the simplest example first. – JavaMan Dec 1 '11 at 18:08
ah, great idea. brb – sasha Dec 1 '11 at 18:25
@David Er, it seems I misread the question. You're right, thanks for pointing it out. I removed the comment now. – Srivatsan Dec 1 '11 at 19:25
This is just an (trivial) observation and may lead to nowhere: draw three tangents on the inscribed circle of radius $r_1$ that are perpendicular to the angle bisectors of the triangle (this is somewhat ill-stated, but I hope the meaning is clear). Then the other three circles are the incircles of the three triangles formed by those tangents and the original triangle. – David Mitra Dec 1 '11 at 19:34
This is an open problem at (that contains a proposed answer)… – David Mitra Dec 1 '11 at 20:13
up vote 4 down vote accepted

I'm using the same notation as in the link mentioned by David Mitra. By similarity we have $$ \frac{r}{OA}=\frac{r_a}{OA-r-r_a}=\sin\left(\frac{\alpha}{2}\right) $$ where $\alpha=\angle A$. Then $$ \frac{r_a}{r} = \frac{1-\sin(\alpha/2)}{1+\sin(\alpha/2)} = \frac{1-\cos((\beta+\gamma)/2)}{1+\cos((\beta+\gamma)/2)} = \tan\left(\frac{\beta+\gamma}{4}\right)^2 $$ where we used that $\alpha+\beta+\gamma=\pi$. Similarly $$ \frac{r_b}{r} = \tan\left(\frac{\alpha+\gamma}{4}\right)^2 \qquad\qquad \frac{r_c}{r} = \tan\left(\frac{\alpha+\beta}{4}\right)^2 $$ Therefore, $\sqrt{r_ar_b}+\sqrt{r_ar_c}+\sqrt{r_br_c}$ can be written as $$ \begin{split} \frac{\sqrt{r_ar_b}+\sqrt{r_ar_c}+\sqrt{r_br_c}}{r} = \frac{\sin(\rho)\sin(\sigma)\cos(\tau)+\sin(\rho)\cos(\sigma)\sin(\tau) +\cos(\rho)\sin(\sigma)\sin(\tau)}{\cos(\rho)\cos(\sigma)\cos(\tau)} \end{split} $$ where $\rho=(\beta+\gamma)/4$, $\sigma=(\alpha+\gamma)/4$, and $\tau=(\alpha+\beta)/4$. Note that $\rho+\sigma+\gamma=\pi/2$. Consider $x$, $y$, $z$ such that $x+y+z=\pi/2$. Then $$ \begin{split} \sin(x)\sin(y)\cos(z) &= \frac{1}{2}(\cos(x-y)-\cos(x+y))\cos(z) \\&= \frac{1}{4}(\cos(x-y+z)+\cos(x-y-z) -\cos(x+y+z)-\cos(x+y-z)) \\&= \frac{1}{4}( \cos\left(\frac{\pi}{2}-2y\right) + \cos\left(-\frac{\pi}{2}+2x\right) - \cos\left(\frac{\pi}{2}\right)-\cos\left(\frac{\pi}{2}-2z\right) \\&=\frac{1}{4}\left( \sin(2y)+\sin(2x)-\sin(2z)\right) \end{split} $$ And similarly $$ \begin{split} \cos(x)\cos(y)\cos(z) &= \frac{1}{2}(\cos(x-y)+\cos(x+y))\cos(z) \\&= \frac{1}{4}(\cos(x-y+z)+\cos(x-y-z) +\cos(x+y+z)+\cos(x+y-z)) \\&=\frac{1}{4}\left( \sin(2y)+\sin(2x)+\sin(2z)\right) \end{split} $$ This means that $(\sqrt{r_ar_b}+\sqrt{r_ar_c}+\sqrt{r_br_c})/r$ is equal to $$ \frac{\sqrt{r_ar_b}+\sqrt{r_ar_c}+\sqrt{r_br_c}}{r} = \frac{\sin(2\rho)+\sin(2\sigma)+\sin(2\tau)}{\sin(2\rho)+\sin(2\sigma)+\sin(2\tau)}=1 $$

share|cite|improve this answer
Wow, this is a great effort--thanks. I've been away from my computer for a few days so I couldn't respond earlier. I have something similar to what you have which I will be posting shortly. – sasha Dec 5 '11 at 17:10

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.