2
$\begingroup$

Let ABCD be a square with side length $a$. Let $M$ be a point on $BC$. The line $AM$ meets the diagonal $BD$ at $K$. From $K$ draw the perpendicular $KL$ to $CD$, with foot $L$. Let $P$ be the reflection of $C$ with respect to $L$. The line $PK$ meets $AB$ at $F$. Prove that the quadrilateral $AFMP$ is an isosceles trapezoid.

If $\frac{BM}{MC}=n$, prove that $FB=\frac{na}{(2n+1)(n+1)}$.

While experimenting with ratios in a square ABCD of side length $a$, I first considered the case where $M$ is the midpoint of $BC$. From a carefully constructed diagram I conjectured that the quadrilateral $AFMP$ should be an isosceles trapezoid, which I then proved synthetically, obtaining $FB=\frac{a}{6}$. This led me to consider a generalization where $M$ is an arbitrary point on $BC$ such that $\frac{BM}{MC}=n$. Using a similar approach, I found that $AFMP$ is again an isosceles trapezoid and that $FB=\frac{na}{(2n+1)(n+1)}$.

I present a synthetic proof below and would be interested in alternative approaches (coordinates, vectors, or transformations), as well as possible generalizations.

enter image description here

Extend $KC$ to meet $AB$ at $T$ and let $KL$ meet $AB$ at $N$. We have $AD=DC$, $DK$ common and $\angle ADK=\angle KDC=90^\circ \Rightarrow \triangle AKD \cong \triangle DKC$ (SAS) $\Rightarrow AK=KC$. Since $L$ is the midpoint of $CP$ and $KL \perp CD$, it follows $KC=KP \Rightarrow AK=KP$ (1).

Also $\angle ABM=\angle CBT=90^\circ$, $AB=BC$, and $\angle BAM=\angle TCB$ (supplementary to $\angle DAK=\angle KCD$) $\Rightarrow \triangle ABM \cong \triangle TBC$ (ASA) $\Rightarrow BT=BM$. Further $\triangle TBK \cong \triangle MBK$ (SAS) $\Rightarrow KT=KM$. Moreover $\triangle TKF \sim \triangle KPC$ (isosceles) $\Rightarrow KT=KF \Rightarrow KM=KF$ (2).

Now $\angle MKF=\angle AKP$ (vertical) and from (1),(2) we get $\triangle AKP \sim \triangle FKM \Rightarrow \angle AMF=\angle MAP \Rightarrow FM \parallel AP$ (3), and $AK+KM=KP+KF \Rightarrow AM=FP$ (4). From (3),(4) it follows that $AFMP$ is an isosceles trapezoid.

For the second part, if $\frac{BM}{MC}=n$, then $\frac{BM}{BC}=\frac{n}{n+1} \Rightarrow BM=\frac{na}{n+1}$ (5). Draw $FZ \perp AB,CD$ with feet $F,Z$. Since $KL \parallel FZ$, by Thales we get $\frac{LZ}{PL}=\frac{KF}{KP}=\frac{KM}{AK}$. Since $BM \parallel AD$, it follows $\frac{KM}{KA}=\frac{BM}{AD}=\frac{BM}{BC}=\frac{n}{n+1}$,

hence $\frac{LZ}{PL}=\frac{TN}{NB}=\frac{n}{n+1} \Rightarrow \frac{NB}{TN}=\frac{n+1}{n} \Rightarrow \frac{NB}{TB}=\frac{NB}{BM}=\frac{n+1}{2n+1} \Rightarrow NB=\frac{na}{2n+1}$ (6).

Since $TN=NF$, we have $NB=\frac{FB+TB}{2}$=$\frac{FB+BM}{2}$, and substituting from (5),(6) we obtain $FB=\frac{na}{(2n+1)(n+1)}$.

$\endgroup$
0

1 Answer 1

2
$\begingroup$

Here is a proof featuring angle chasing.

$$\triangle AKB\cong\triangle CKB\implies \angle KAB=\angle KCB=90^\circ-\angle AMB=\angle LKC=\angle LKP=90^\circ-\angle KPC$$ implies that $KPCM$ is cyclic. Also, $\angle MAF=\angle MPF$ implies $AFMP$ is also cyclic. Since $AK=KC=KP$, $\angle PAK=\angle APK=\angle PFM$ which concludes that $AFMP$ is a trapezium. But $AFMP$ is cyclic which is possible iff it is an isosceles trapezium.

$\endgroup$
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.