MathOverflow is a question and answer site for professional mathematicians. 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

Is it true that of all $n$-simplices with edge lengths greater than or equal to some parameter $s$, the regular simplex with edge lengths $s$ has the smallest circumradius? It seems obvious, but I haven't been able to prove this.

I would greatly appreciate if someone could supply a proof or reference. Thank you.

share|cite|improve this question
Wow! Two hours and still no answer?! I better play MO's savior :-) – Włodzimierz Holsztyński May 15 '13 at 3:08
up vote 4 down vote accepted

Let   $H$   be a Hilbert space. Let   $e_1\ \ldots\ e_n$   be vectors such that   $\forall_{k=1\ldots n}\ \|e_k\|\le 1$   for a natural   $n>1$.  I am addressing $(n-1)$-simplex for the sake of avoiding eyesores. Then:

$$ 0\ \le\ (e_1+\ldots +e_n)^2\ \ \le\ \ n\ +\ 2\cdot\sum_{j\ne k}e_j\cdot e_k $$

It follows that there exist   $j\ne k$   such that

$$ e_j\cdot e_k\ \ \ge\ \ \frac{-n}{2\cdot\binom n2}\ =\ \frac1{1-n}$$


$$\|e_j-e_k\|\ =\ \sqrt{e_j^2 - 2\cdot e_j\cdot e_k + e_k^2}\ \le\ \sqrt{2\cdot(1+\frac 1{n-1})}\ =\ \sqrt{\frac{2\cdot n}{n-1}}$$


$$\|e_j-e_k\|\ \le\ \sqrt{\frac{2\cdot n}{n-1}}$$

REMARK 0   We get equality above   $\Leftrightarrow$   $\|e_j\|=\|e_k\|=1$   and   the   $\binom n2$   values   $e_j\cdot e_k$   are all equal one to another, i.e.   all distances   $\|e_j-e_k\|$   are all equal. (Then of course all dot products are equal to   $\frac 1{1-n}$,   and the distances to   $\sqrt{\frac{2\cdot n}{n-1}}$).

Now let's reverse our point of view. Let   $e_1\ldots e_k\in H$   be such that   $\|e_j-e_k\|\ge s$   for every   $j\ \,k = 1\ldots n$   such that   $j\ne k$,   where   $s>0$   is an arbitrary positive real constant. Furthermore, we may assume that the center of a sphere which has   $e_1\ldots e_k$   for its points coincides with the origin of   $H$,   so that   $\|e_1\|=\ldots =\|e_n|=r$   for certain positive real   $r$.   Then

$$ s\ \ \le\ \ r\cdot \sqrt{\frac{2\cdot n}{n-1}}$$


$$ r\ \ \ge\ \ s\cdot\sqrt{\frac{n-1}{2\cdot n}}$$

The equality holds   $\Leftrightarrow$   all edges have the same length   $\|e_j-e_k\|=s$,   and vectors   $e_1\ldots e_n$   have to belong to a common $(n-1)$-dimensional linear subspace of   $H$   (substitute linear by affine in the general case).

share|cite|improve this answer
Thank you very much for your answer! This is a bit embarrassing, but can I ask you to supply a bit more detail? In particular, where does the circumradius come into play? – user21277 May 15 '13 at 4:15
You ask for the shortest radius when the edges are long. I look above for the longest minimum of edges when the radius is short. It leads to your question. It's a standard duality for such situations. (And you can always move your simplex to position the center of the sphere at the origin of the space; I'll expand the text of my "Answer" a little later; first let my eyes rest from the $\LaTeX$ struggles and battles from other MO questions. – Włodzimierz Holsztyński May 15 '13 at 5:51
Edges are intervals $(e_j\ e_k)$ for $j\ne k$. – Włodzimierz Holsztyński May 15 '13 at 5:53
Thanks a lot! It's a very clean solution. I was trying to use the Cayley-Menger determinant and was having no success. It's surprising to me that the solution avoids such complications. – user21277 May 15 '13 at 6:48
Otherwise (if it were complex) I would not be able to provide it :-). You're most welcome. BTW. I introduced this method (most likely very well known?) for some other geometric results in the past. – Włodzimierz Holsztyński May 15 '13 at 6:57

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.