We have given isosceles triangle $ABC$ with baseline $AB$. Point $M$ is midpoint of AB. Draw perpendicular line to side $AC$ through $M$ which intersects $AC$ in point $H$. Let $P$ be midpoint of $MH$. Show that lines $BH$ and $CP$ are perpendicular to each other.

I also have "hint": Construct height $BD$ to side $AC$ and look at the triangles $ABD$ and $MCH$.

Analytically there is no problem but I get stuck in geometrical approach to solution.


We draw the line $BD$ perpendicular to $AC$ and parallel to $MH$. Because $\angle CBA=\angle CAB=\angle CMH$, we know that $\triangle AMH\sim\triangle ABD\sim\triangle MCH$. Because $|AM|=|MB|$, it follows that $|AH|=|HD|$. Thus, $H$ is the midpoint of $AD$. Now, we see that $BH$ and $CP$ are medians in $\triangle ABD$ and $\triangle MCH$. Also, we can get $\triangle MCH$ by rotating $\triangle ABD$ over $90^\circ$ and translating and scaling it. Because translation and scaling don't change the angle between corresponding lines, we now find that $BH\perp CP$.


  • 1
    $\begingroup$ Thanks for detailed explanation + thumbs up for geogebra :) $\endgroup$ – markich Jan 5 '14 at 21:51

I hope the following can satisfy the requirement of a geometrical solution.

I modified Ragnar’s drawing as follow

As pointed out:-

(1) HD = (1/2) AD ------------------------ [intercept theorem]

(2) △ABD ∼ △MCH-------------------- [equi-angular]

Then, from (2), AD / BD = MH / CH ----------- (*)

In △HCP, tan y = HP / CH

Then, tan y = (1/2)(MH) / CH -------------------- [given]

In △HBD, tan z = HD / BD

Then tan z = (1/2)(AD) / BD --------------------- [from (1)]

From (*), we have tan y = tan z

Hence, y = z [since none of the angles are more than π.]

Then, C, D, K and B are con-cyclic ---------[converse of angles in the same segment]

angle CKB = angle CDB = 90 degree------ [angles in the same segment]

  • $\begingroup$ Please take a look on the following question: math.stackexchange.com/questions/2321816/… $\endgroup$ – Blind Jun 15 '17 at 21:15
  • $\begingroup$ @Blind I have seen your question when it was posted. It looks a bit complicated but I will take a second look at it. Just wondering, is that question somehow related to this answer? $\endgroup$ – Mick Jun 16 '17 at 1:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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