Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

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

Is there any geometrically simple reason why the inradius of a triangle should be at most half its circumradius? I end up wanting the fact for this answer.

I know of two proofs of this fact.

Proof 1:

The radius of the nine-point circle is half the circumradius. Feuerbach's theorem states that the incircle is internally tangent to the nine-point circle, and hence has a smaller radius.

Proof 2:

The Steiner inellipse is the inconic with the largest area. The Steiner circumellipse is the circumconic with the smallest area, and has 4 times the area of the Steiner inellipse. Hence the circumcircle has at least 4 times the area of the incircle.

These both feel kind of sledgehammerish to me; I'd be happier if there were some nice Euclidean-geometry proof (or a way to convince myself that no such thing is likely to exist, so the sledgehammer is necessary).

EDIT for ease of future searching: The internet tells me this is often known as "Euler's triangle inequality."

share|cite|improve this question
up vote 10 down vote accepted

So Proof #1 can be modified to be completely elementary.

First, it is easy to show that the incircle is the smallest circle touching all 3 sides. The circle passing through the midpoints (the nine-point circle) obviously has circumradius half that of the larger circle. No need to invoke Feuerbach's theorem for this.



share|cite|improve this answer
Good point. Thanks! – Micah Jul 14 '12 at 21:20

Compute the area of a triangle (first method):

Consider the following diagram:

$\hspace{4.5cm}$enter image description here

The area of the green triangle is $$ A=\tfrac12ab\sin(\theta)\tag{1} $$ By the Inscribed Angle Theorem, the angle that $c$ subtends at the origin, $o$, is $2\theta$. Therefore, we get that $$ c=2R\sin(\theta)\tag{2} $$ Combining $(1)$ and $(2)$ yields $$ 4AR=abc\tag{3} $$ Compute the area of a triangle (second method):

Consider the following diagram:

$\hspace{4.5cm}$enter image description here

Note that the areas of the the red ($\color{#C00000}{\triangle iyz}$), green ($\color{#00A000}{\triangle izx}$), and blue ($\color{#0000FF}{\triangle ixy}$) triangles are $\frac12r$ times $a$, $b$, and $c$, respectively. Therefore, we get $$ 2A=r(a+b+c)\tag{4} $$

Compute $d$:

Translate the circumcenter, $o$, of $\triangle xyz$ to the origin. Then $$ |x|=|y|=|z|=R\tag{5} $$ Furthermore, using $a=|y-z|$, $b=|z-x|$, and $c=|x-y|$, we get $$ \begin{align} 2y\cdot z&=2R^2-a^2\\ 2z\cdot x&=2R^2-b^2\\ 2x\cdot y&=2R^2-c^2 \end{align}\tag{6} $$ As mentioned above, the areas of the the red ($\color{#C00000}{\triangle iyz}$), green ($\color{#00A000}{\triangle izx}$), and blue ($\color{#0000FF}{\triangle ixy}$) triangles are proportional to $a$, $b$, and $c$, respectively. Thus, the barycentric coordinates of the incenter, $i$, are the mean of the vertices weighted by the lengths of the opposite sides: $$ i=\frac{ax+by+cz}{a+b+c}\tag{7} $$ and therefore, using $(3)$-$(7)$ yields that $d$, the distance between the incenter and circumcenter, satisfies $$ \begin{align} d^2 &=\frac{a^2R^2+b^2R^2+c^2R^2+2abx\cdot y+2bcy\cdot z+2caz\cdot x}{(a+b+c)^2}\\ &=\frac{a^2R^2+b^2R^2+c^2R^2+ab(2R^2-c^2)+bc(2R^2-a^2)+ca(2R^2-b^2)}{(a+b+c)^2}\\ &=\frac{(a+b+c)^2R^2-(a+b+c)abc}{(a+b+c)^2}\\ &=R^2-\frac{abc}{a+b+c}\\[4pt] &=R^2-2Rr\\[6pt] &=R(R-2r)\tag{8} \end{align} $$ Since $d^2\ge0$, and $R>0$, we immediately get that $$ r\le\tfrac12R $$

share|cite|improve this answer
Thanks! I think this might strictly speaking be overkill as it's a lot more complex than the translational argument that Rofler didn't give, but it's great to have an expression for exactly how unequal the two quantities are... – Micah Jul 16 '12 at 20:15
@Micah: I moved this part of my answer to this question since it applies more directly here. I reference it from the other answer. – robjohn Jul 16 '12 at 20:47
@Micah: I didn't start from the 9-point circle. I hope this is simpler when starting from scratch. – robjohn Jul 16 '12 at 20:55
To my mind the other argument is still simpler -- note that you don't actually need to know that it's the 9-point circle, just that it intersects the three midpoints. But I guess in some sense it's a question of algebraic vs. geometric complexity; visualizing the other proof is easier, but writing down an explicit set of translations that does the trick would probably require more work than what you did here. – Micah Jul 16 '12 at 21:30
@Micah: Okay; I see. If we characterize the incircle as the smallest circle that touches all three sides, then yes, the circumcircle of the triangle constructed from the midpoints of the sides would be bigger than the incircle, and it would have half the radius of the original circumcircle. – robjohn Jul 17 '12 at 0:33

Expanding on Rofler's answer, just so there'll be a complete argument here:

Consider the circle which meets the midpoints of the three sides (this is secretly the 9-point circle, but we don't need to know that to finish this argument). It circumscribes a triangle which is similar to the reference triangle and scaled by $\frac{1}{2}$. Thus its radius is half the circumradius.

To show that the radius of the nine-point circle is larger than the inradius of the reference triangle, first translate it in your favorite direction until the first time it becomes tangent to a triangle side. Then slide it along that side of the triangle until the first time it becomes tangent to another triangle side. The result will be a circle which is tangent to two sides of the reference triangle and intersects the third, and any such circle must clearly be at least as large as the incircle.

share|cite|improve this answer

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.