An ellipse with the center at $(3,4)$ touches the $x$-axis at $(0,0)$. If the slope of its major axis is $1$, find the eccentricity and equation of the ellipse.

What I try:

Let major axis be $x-y=c$ and minor axis be $x+y=d$. Then $\displaystyle x=\frac{d+c}{2}$ and $\displaystyle y=\frac{d-c}{2}$

So we have $\displaystyle =\frac{d+c}{2}=3$ and $\displaystyle \frac{d-c}{2}=4$

So $\displaystyle d+c=6$ and $d-c=8,$ Solving $d=7,c=-1$

major axis equation is $\displaystyle x-y+1=0$ and minor axis equation is $x+y-7=0$

How do I solve it? Help me, please.

| cite | improve this question | | | | |

You have correctly identified the axes. So the equation of the ellipse will be $$\frac{(x+y-7)^2}{a^2}+\frac{(x-y+1)^2}{b^2}=1$$ for some $a,b$. [We have effectively changed coordinates from $x,y$ to $x+y,x-y$.]

We know that it passes through the origin so we need $\frac{49}{a^2}+\frac{7}{b^2}=1$.

Assume $x,y$ are functions of some parameter (eg set $x=a\cos t,y=b\sin t$). Differentiate the equation of the ellipse and set $x=y=\dot{y}=0$ and we get $$\frac{2}{a^2}(-7)(\dot{x})+\frac{2}{b^2}(1)(\dot{x})=0$$ or $a^2=7b^2$. The two equations for $a^2,b^2$ give $a^2=56.b^2=8$.So the equation of the ellipse is:

enter image description here

You can now get the eccentricity from $$e=\sqrt{1-\frac{b^2}{a^2}}=\sqrt{\frac{6}{7}}$$

| cite | improve this answer | | | | |
  • $\begingroup$ A very elegant solution :-) $\endgroup$ – Lakshya Sinha Jan 2 at 14:40

In a modification of the answer by almagest, I explore finding a solution without theuse of differentiation.

My starting point in the first equation derived by almagest:


In this approach put in $y =0$ and derive a quadratic equation for $x$:



To get a tangent at $(0,0)$ this equation must have a double root at $x=0$, forcing the linear and constant coefficients both to be zero:

$14a^2-2b^2=0$ Eq. 1

$49a^2+b^2-a^2b^2=0$ Eq. 2

Eq. 1 renders $b^2=7a^2$ and plugging that into Eq. 2 then gives $a^2=8$. Thus


as found by almagest with his method.

| cite | improve this answer | | | | |
  • 1
    $\begingroup$ +1 Yes,you are right. It is probably better to avoid calculus. $\endgroup$ – almagest Jan 2 at 14:50
  • 1
    $\begingroup$ +1 indeed, Not trying to be picky but Eq 1. gives $14a^2$ :p you missed out 2 in $2b^2$ ;p $\endgroup$ – Chief VS Jan 2 at 14:54
  • 1
    $\begingroup$ Thanks @chief. That is officially my millionth typo on this site. Cheers! :-S $\endgroup$ – Oscar Lanzi Jan 2 at 15:04

Here’s another way to use the tangency constraint that doesn’t involve calculus. With what you’ve found so far, you know that the equation of the ellipse has the form $${(x+y-7)^2\over a^2}+{(x-y+1)^2\over b^2} = 1.\tag1$$ Note that the equation of the minor axis goes with the major semiaxis length, and vice-versa. The tangent to this ellipse at the point $(x_0,y_0)$ has the equation $${(x+y-7)(x_0+y_0-7)\over a^2}+{(x-y+1)(x_0-y_0+1)\over b^2} = 1.\tag 2$$ Setting $x_0=y_0=0$ in equation (2) and requiring that the coefficient of $x$ vanish (the tangent at the origin has an equation equivalent to $y=0$) produces the equation $${a^2-7b^2\over a^2b^2}=0,\tag3$$ from which $a^2=7b^2$. We can immediately compute the eccentricity: $$\sqrt{1-b^2/a^2}=\sqrt{1-1/7}=\sqrt{6/7}.$$ It’s important to note that this only works because the normals implied by the equations of the principal axes have the same length. This ensures that even though the coefficients $a$ and $b$ in equation (1) aren’t equal to the actual semiaxis lengths of the ellipse, they’re in the same proportion to those lengths.

To find $a^2$ and $b^2$, set $x=y=0$ in equation (1) and use $a^2=7b^2$ to obtain $a^2=56$ and $b^2=8$.

| cite | improve this answer | | | | |

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.