2
$\begingroup$

Let $ABCD$ be a quadrilateral. Do there exist points $P,Q,R,S$ such that $A,B,C$ and $D$ are the respective circumcenters of $\triangle QRS$, $\triangle RSP$, $\triangle SPQ$ and $\triangle PQR$?

Clearly, three such points always exist in a plane so proof is complete if the fourth point also exists. I tried using radical axis from which I got $AB\perp RS,BC\perp SP,CD\perp PQ$ so essentially proving $AD\perp QR$ is enough. But I am not able to make a figure.

Progress: Assume the fourth point (WLOG let it be $Q$) exists. Note that the line passing through midpoint of perpendicular bisector of $QS$ will pass through $A$ and $C$, in other words, $A,C$ and the midpoint of radical axis of circle with center $A,C$ ($QS$) will be collinear. Similarly, consider the midpoints of $RS,SP,PQ,QR$ to be $W,X,Y,Z$. So, collinear points are $$A,B,W$$

$$B,C,X$$

$$C,D,Y$$

$$D,A,Z$$

Figure

Therefore, proving that $Y,Z$ as not fixed and $W,X$ as fixed points will give a contradiction

$\endgroup$

1 Answer 1

1
$\begingroup$

Such points $P$, $Q$, $R$, $S$ do not always exist.


To see this, consider the (arguably) simplest quadrilateral - the unit square. We place this in coordinate space with $A(0,0)$, $B(1,0)$, $C(1,1)$, $D(0,1)$.

If $A$ is the circumcenter of $\triangle QRS$, then $A$ is equidistant from $Q$, $R$, and $S$. Similarly, $B$ is equidistant from $R$, $S$, and $P$. Since $A$, $B$ are both equidistant from $R$, $S$, we have that $AB$ is the perpendicular bisector of $RS$. Hence, $x_R=x_S$ and $y_R+y_S=0$.

Similarly, we have $x_S+x_P=2$, $y_P+y_Q=2$, and $x_Q+x_R=0$.

Define $x_R=x$. Then, $x_P=2-x$, $x_Q=-x$, and $x_S=x$. Also, define $y_R=y$. This forces $y_S=-y$.

Since $B$ is equidistant from $R$ and $P$, $$(x-1)^2+y^2=(1-x)^2+y_P^2 \implies y_P= \pm y$$

Since $C$ is equidistant from $S$ and $P$, $$(x-1)^2+(-y-1)^2=(x_P-1)^2+(y_P-1)^2$$ Since $x_P=2-x$, $$(x-1)^2+(-y-1)^2=(1-x)^2+(y_P-1)^2 \implies y_P=-y \text{ or } y_P=y+2.$$

So, either $y=0$ (degenerate case since $R$ and $S$ would coincide) or $y_P=-y$.

Since $D$ is equidistant from $R$ and $Q$, $$x^2+(y-1)^2=x_Q^2+(y_Q-1)^2$$ Since $x_Q=-x$, $$(y-1)^2=(y_Q-1)^2 \implies y_Q=y \text{ or } 2-y.$$

However, we have from earlier that $y_P+y_Q=2$. Since $y_P=-y$, $y_Q=2+y$ regardless. But then the only possibility is for $y=0$, which we already established to lead to a degenerate case.

$\endgroup$
0

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.