2 tangents to the parabola $y^2=4ax$ meet at an angle of $45^\circ$. Prove that the locus of their point of intersection is $y^2-4ax=(x+a)^2$.

$$$$ I got completely stuck with this question. All I could think of was that the tangents would be of the form $$y=m_1x-am_1^3-2am_1$$ and $$y=m_2x-am_2^3-2am_2$$ where $m_1$ and $m_2$ are the slopes of the respective tangents. I also know that $\tan 45=|\dfrac{m_1-m_2}{1-m_1m_2}|$ $$$$ Using this, I tried to work out the value of $m_2$ in terms of $m_1$. However I got stuck because there were 2 possible values of $m_2$ that I got from the $\tan 45$ equation, and I couldn't decide which value of $m_2$ to use. $$$$Lastly, I know the result that if there are 2 tangents at $t_1$ and $t_2$, their point of intersection is at $(at_1t_2, a(t_1+t_2)$

However, I cannot understand how to apply these results, or to use some other method. I would be truly grateful if someone could please help me with this problem. Many thanks!


the tangents would be of the form $$y=m_1x-am_1^3-2am_1$$ and $$y=m_2x-am_2^3-2am_2$$ where $m_1$ and $m_2$ are the slopes of the respective tangents.

I think that this is not correct.

Let $(s,t)$ be a point on the parabola. Then, since $m=\frac{2a}{t}\Rightarrow t=\frac{2a}{m}$ with $t^2=4as$, $$y-t=\frac{2a}{t}(x-s)\iff y-\frac{2a}{m}=m\left(x-\frac{1}{4a}\left(\frac{2a}{m}\right)^2\right)$$So, $$y=m_1x+\frac{a}{m_1},\quad y=m_2x+\frac{a}{m_2}$$

I also know that $\tan 45=|\dfrac{m_1-m_2}{1-m_1m_2}|$

Note that we have $$\tan(45^\circ)=\left|\frac{m_1-m_2}{1\color{red}{+}m_1m_2}\right|\tag1$$

Since the intersection point $(X,Y)$ is $\left(\frac{a}{m_1m_2},\frac{m_1+m_2}{m_1m_2}a\right)$, we have $$m_1m_2=\frac{a}{X},\quad m_1+m_2=\frac YX.$$

Hence, from $(1)$ $$(1+m_1m_2)^2=(m_1-m_2)^2=(m_1+m_2)^2-4m_1m_2$$ $$\left(1+\frac aX\right)^2=\left(\frac YX\right)^2-4\cdot\frac aX,$$ i.e. $$Y^2-4aX=(X+a)^2$$

  • $\begingroup$ I'm really sorry. I got confused with the equations of tangents and normals. Really sorry once again. Thanks very much for your help and patience. $\endgroup$ – Better World Jan 7 '16 at 8:22
  • $\begingroup$ Sir, could you please explain why you've equated with the expression inside the modulus with $1$ instead of $\pm 1$? $\endgroup$ – Better World Jan 7 '16 at 8:27
  • $\begingroup$ @BetterWorld I think he has equated the square of that expression to $1$ which is perfectly fine. $\endgroup$ – G-man Jan 7 '16 at 8:28
  • $\begingroup$ But Sir, if we have say line $l_1$ with slope $m_1$ which is known, then there are 2 possible lines $l_2$ and $l_3$ with slope $m_2$ and $m_3$ respectively which could make a $45^\circ$ angle with $l_1$. However, from $\tan(45^\circ)=\frac{m_1-m_2}{1+m_1m_2}$ we woulcdd get only one value of $m$ which could be the value of $m_2$ or $m_3$. This, we would get the slope of only one of the 2 possible lines. Isn't that why we use $\tan(45^\circ)=\tan(45^\circ)=\left|\frac{m_1-m_2}{1+m_1m_2}\right|$, to get both the slopes of both possible lines $l_2$ and $l_3$? $\endgroup$ – Better World Jan 7 '16 at 10:17
  • $\begingroup$ @BetterWorld: Sorry, forget my last comment (which I already deleted). We have $(1)$ in my answer, which is $1=\left|\frac{m_1-m_2}{1+m_1m_2}\right|$. So, we have $|1+m_1m_2|=|m_1-m_2|$. Therefore, we have $(1+m_1m_2)^2=(m_1-m_2)^2$. Any confusion about these steps? $\endgroup$ – mathlove Jan 7 '16 at 10:25

Corresponding to parameter value $t$, the gradient of the tangent is $\frac 1t$.

Therefore, from the angle formula, you have $$\left|\frac{\frac{1}{t_2}-\frac{1}{t_1}}{1+\frac{1}{t_1t_2}}\right|=\left|\frac{t_1-t_2}{1+t_1t_2}\right|=1$$

You also have $\frac xa=t_1t_2$ and $\frac ya=t_1+t_2$

Writing the first equation as $$|t_1-t_2|=|1+\frac xa|$$ and squaring both sides, we get $$\left(1+\frac xa\right)^2=(t_1-t_2)^2=(t_1+t_2)^2-4t_1t_2=\left(\frac ya\right)^2-4\frac xa$$

Hence the result follows immediately

  • $\begingroup$ Sir, could you please explain why $\dfrac{t_1-t_2}{1+t_1t_2}=1$ instead of $\dfrac{t_1-t_2}{1+t_1t_2}=\pm1$? $\endgroup$ – Better World Jan 7 '16 at 8:23
  • $\begingroup$ sorry, I have edited my answer! $\endgroup$ – David Quinn Jan 7 '16 at 8:36

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.