Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

In class, we were simply given that this limit is undefined since along the paths $y=\pm x$, the function is undefined.

Am I right to think that this should be the case for any function, where the denominator is $x^2-y^2$, regardless of what the numerator is?

Just wanted to see if this is a quick way to identify limits of this form.

Thanks for the discussion and help!

share|cite|improve this question
up vote 10 down vote accepted

If you define $$\lim_{\langle x,y\rangle\to\langle a,b\rangle}f(x,y)\tag{1}$$ in such a way that it exists only when the function is defined in some open ball centred at $\langle a,b\rangle$, then what you wrote is correct. This is analogous to defining $$\lim_{x\to a}f(x)$$ only when $f(x)$ is defined in some open interval centred at $a$. However, just as we can talk about one-sided limits on the real line, it makes perfectly good sense to talk about $(1)$ whenever $f(x,y)$ is defined at points arbitrarily close to $\langle a,b\rangle$. In that case it’s understood that we look only at the limit along ‘paths’ within the domain of $f$. On that understanding $$\lim_{\langle x,y\rangle\to\langle 0,0\rangle}\frac{x^3+y^3}{x^2-y^2}$$ still does not exist, but for a more fundamental reason.

Suppose that you approach the origin along the curve $y=\sin x$. Then by l’Hospital’s rule you have

$$\begin{align*} \lim_{\langle x,y\rangle\to\langle 0,0\rangle}\frac{x^3+\sin^3x}{x^2-\sin^2x}&=\lim_{\langle x,y\rangle\to\langle 0,0\rangle}\frac{3x^2+3\sin^2x\cos x}{2x-2\sin x\cos x}\\\\ &=\lim_{\langle x,y\rangle\to\langle 0,0\rangle}\frac{3x^2+\frac32\sin2x\cos x}{2x-\sin2x}\\\\ &=\lim_{\langle x,y\rangle\to\langle 0,0\rangle}\frac{6x-\frac32\sin2x\sin x+3\cos2x\cos x}{2-2\cos2x} \end{align*}$$

which does not exist: the numerator approaches $3$ and the denominator, $0$. The problem is that this path, although it stays within the domain of the function, approaches the line $y=x$ so quickly as it approaches the origin that the denominator approaches $0$ much faster than the numerator, and therefore the function blows up as we approach the origin along this path.

share|cite|improve this answer
@DayLateDon: Ouch. I sure do; thanks! – Brian M. Scott Jun 3 '12 at 8:16
Compelet answer. :-) – Babak S. Jun 3 '12 at 8:35

Brian's answer is nice, so it got me thinking, what about approaching the point along other functions, $y=f(x)$, with $f(0)=0$. Let us plug that into the expression to get $\frac{x^3+f^3}{x^2-f^2}$. This will give $$\lim_{x \to 0} \frac{x^2-xf+f^2}{x-f}\;.$$

We will now perform a l’Hospital’s rule to get $$\lim_{x \to 0} \frac{2x-xf'-f+2ff'}{1-f'}\;.$$ If at this point we assume that $f'(0)=1$, we have another $0/0$. If $f'\neq 1$, we get that the limit is zero. So letus now hit this with another l’Hospital. This will give us $$\lim_{x \to 0} \frac{2-2f'-xf''+2f''f'+2(f')^2}{-f''}\;.$$

When evaluate at $x=0, f'=1$, we get $$\lim_{x \to 0} -\frac{2f''+2}{f''}\;.$$ And you can decide what you want the limit to be by picking a value for $f''$.

share|cite|improve this answer
I fixed an obvious typo (you had ‘If $f'\ne0$’ instead of $f'\ne1$) and moved the periods inside the displayed formulas. – Brian M. Scott Jun 3 '12 at 8:22
Thank you. =))) – Baby Dragon Jun 3 '12 at 8:29

It doesn't matter that along the lines $y = \pm x$, the function is undefined ; you can still stay inside the domain of $f(x,y) = \frac{x^3 + y^3}{x^2-y^2}$ and try to compute the limit, under the constraint that $y \neq \pm x$. The limit could still be defined at $(0,0)$.

But there are cases where your trick doesn't work : think of $\frac{x^4-y^4}{x^2-y^2} = x^2 + y^2$, which is continuous at $(0,0)$.

Hope that helps,

share|cite|improve this answer

For your function, in the domain of $f$ (so $x\ne \pm y)$, to compute the limit you can set $x=r\cos\theta, y=r\sin\theta$, and plug it it. You get $\lim\limits_{r\to 0} \frac{r^3(cos^3\theta+sin^3\theta)}{r^2(cos^2\theta-sin^2\theta)} =\lim\limits_{r\to 0} \frac{r(cos^3\theta+sin^3\theta)}{(cos^2\theta-sin^2\theta)}$, and you can easily see that this is $0$ for any $\theta$ in the domain of $f$ (you need to avoid $\theta = \frac{n\pi}{2}-\pi/4$). Of course, if you are considering the whole plane, then the limit does not exist, because the function isn't even defined at $y=x$, so you can't compute the limit along that path.

share|cite|improve this answer
There are paths within the domain of $f$ along which the limit is not $0$; e.g., $\theta(t)=\tan^{-1}\left(\frac{\sin t}t\right)$ and $r(t)=\sqrt{t^2+\sin^2t}$ as $t\to 0$. – Brian M. Scott Jun 3 '12 at 6:59
My mistake. Also you need $\theta \ne \frac{n\pi}{2}-\pi/4$. I edited my post. – JLA Jun 4 '12 at 18:21
It’s still wrong, I’m afraid. As @Baby Dragon and I showed, there are paths in the domain along which the limit is not $0$. Your calculation is valid only for straight-line paths. – Brian M. Scott Jun 4 '12 at 20:55

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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