I am doing the end of chapter questions on parabolas in a pure maths book out of interest. I am struggling with this:

A fixed point $P(ap^2,2ap)$ is taken on the parabola $y^2=4ax$. Two points Q and R are chosen on the parabola such that the lines PQ and PR are perpendicular. Prove that the line QR passes through a fixed point, F, independent of Q and R, and that PF is normal to the parabola at P.

I started by working out the slopes of PQ,QR and PR:

$(\frac{2}{p+q},\frac{2}{r+q},\frac{-(p+q)}{2})$ and then I worked out the equation of QR, which I found to be:

$y-2aq=\frac{2}{r+q}(x-aq^2) \rightarrow y=\frac{2x+2aqr}{r+q}$

The question wants an answer independent of Q and R so I reasoned I need to find an expression in p only. I did find such an expression but even then I couldn't find where the fixed point should be. But even then, the question only gives the fixed point F after we are supposed to have found that such a point exists.

enter image description here

  • 1
    $\begingroup$ That $PF$ is the normal at $P$ can be proved independently: in the limit $Q\to P$ line $PQ$ is tangent and chord $QR=PR$ is by construction perpendicular to it. $\endgroup$ Aug 20 at 19:32

Point $P$ on the parabola is $(ap^2, 2ap)$. Say, coordinates of $Q$ and $R$ are $(aq^2, 2aq)$ and $(ar^2, 2ar)$.

Equation of line passing through $QR$,

$y - 2aq = \cfrac{2aq - 2ar}{aq^2-ar^2} (x - a q^2)$


$ (q + r) y = 2x + 2 a q r \tag1$

As $PQ$ and $PR$ are perpendicular,

$\cfrac{2ap - 2aq}{ap^2 - aq^2} \times \cfrac{2ap - 2ar}{ap^2 - ar^2} = - 1$

$\implies \cfrac{4}{(p+q)(p+r)} = - 1$

Simplifying, $p^2 + (q+r) p + 4 + q r = 0$

Multiply by $- 2a$ and rearrange to bring it in the same form as $(1)$.

$ - 2 a p (q+r) = 2 a (p^2 + 4) + 2 a q r \tag2$

From $(1)$ and $(2)$, we can see that $x = a (p^2 + 4), y = - 2 a p$ satisfies the equation of the line $QR$. So, regardless of the value of $q$ and $r$, point $F \big(a (p^2 + 4), - 2 a p \big)$ lies on $QR$.

Equation of normal through point $P$ is,

$y - 2 ap = - p (x - ap^2)$

and you can check that $F$ lies on the normal through $P$.

  • 1
    $\begingroup$ For anyone who's interested, I worked on a Desmos graph when I tried to solve this question a while ago. Thanks to MathLover's answer and something I completely missed, the graph is now complete, and you can visualise the fixed point if you wish (playing with the graph at first I was convinced a fixed point didn't exist, but it all checks out): here $\endgroup$
    – FShrike
    Aug 20 at 18:04
  • 1
    $\begingroup$ @FShrike thank you. that's a good interactive graph. On the contrary, when the question said that the fixed point lied on the normal, I had an intuition it would be circumcenter of the right angled triangle $ \triangle PQR$ but it became difficult to use that information in the working. I finally had to abandon that approach :) $\endgroup$
    – Math Lover
    Aug 20 at 18:20
  • $\begingroup$ and it took me a while to see that my intuition was wrong. $\endgroup$
    – Math Lover
    Aug 20 at 18:34
  • 1
    $\begingroup$ You're welcome. I found the quadratic equation right off, and noticed that it looked very similar to the line equation, but missed the multiplication entirely. Good job $\endgroup$
    – FShrike
    Aug 20 at 18:43
  • 1
    $\begingroup$ It was a very good job indeed. I too explored the fact that PQ is the diameter of a circle but got nowhere. I certainly would never had done it without the help of MathLover. $\endgroup$
    – Steblo
    Aug 20 at 19:33

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.