Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

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

This question is arising from the answer to another one: How find this equation integer solution: $x^2y^2=4x^5+y^3$ . For $x < 27$ and $y > -243$ , the basic equation $x^2 y^2 = 4 x^5 + y^3$ is a function. By implicit differentiation we have found that: $$ y'(x) = \frac{20 x^4 - 2 x y^2}{2 y x^2 - 3 y^2} \quad \Longrightarrow \quad y'(0) = \; ? $$ From the picture it is suspected that: $y'(0) = 0$ , i.e. the slope of the tangent in $(x,y) = (0,0)$ may be zero. But I could not prove or disprove it. Any ideas?

When solving for $y$ (with help of MAPLE) we find something that looks like a decent function, within a prescribed range e.g. $-1 < x < +2$ ; see picture. (Bonus: integer solutions original question at red spots) $$ y(x) = \left[\frac{\left(-54 x^2 + x^3 + 6 \sqrt{81 x^4 - 3 x^5}\right)^{1/3}}{3} + \frac{x^2}{3\left(-54 x^2 + x^3 + 6 \sqrt{81 x^4 - 3 x^5}\right)^{1/3}} + \frac{x}{3}\right] x $$ So it's still not clear to me why the derivative $y'(0)$ would be somehow undefined.

enter image description here     enter image description here

A rather extreme close-up , namely $-1/10 < x < 1/10$ : picture on the right , doesn't reveal any other slope than zero at $(0,0)$ . Whiter means that the function $f(x,y) = 4 x^5 + y^3 - x^2 y^2\;$ is closer to zero; it is seen that $\;f(x,y)\;$ is very close to zero indeed in the neighborhood of $(0,0)$ , thus suggesting that the tangent may be ambiguous there. But is it?

My try.   Draw a circle with radius $r > 0$ and $(0,0)$ as its midpoint: $$ x = r \cos(\phi) \qquad y = r \sin(\phi) $$ Substitute this into the basic equation $\;x^2 y ^2 = 4 x^5 + y^3\;$ and divide by $r^3$ : $$ 4 \cos^5(\phi)\, r^2 - \cos^2(\phi) \sin^2(\phi)\, r + \sin^3(\phi) = 0 $$ If $\;r \rightarrow 0\;$ i.e. becomes very small, then function values in the neighborhood of $\;(0,0)\;$ only depend on the last term $\;\sin^3(\phi)$ . Meaning that $\phi \approx 0$ or $\phi \approx \pi$ . The tangent through these two points has slope zero. Don't know if this counts as a proof.

share|cite|improve this question
I do not believe the derivative is defined at $(0, 0)$. – apnorton Feb 1 '14 at 19:59
Since the equation has no linear terms, anorton is right that the curve does not have a well-defined tangent line at the origin. On the other hand, the tangent cone of the curve at the origin is cut out by the equation $y^3 = 0$, a horizontal line, which is where your suspicion that the slope of the tangent line is zero comes from. – JHF Feb 2 '14 at 0:44
See my update to the question: I still cannot believe that the derivative is undefined at $(0,0)$, but MAPLE doesn't help me as well, with the latter issue. – Han de Bruijn Feb 2 '14 at 13:46
up vote 1 down vote accepted

All the issues here occur also with the simpler equation $x^5=y^3$. Indeed, just as the solutions to that equation can be parametrized as $(x,y) = (t^3,t^5)$, the solutions to the equation you ask about can be parametrized as $(t^3/(4-t)^2, - t^5/(4-t)^3)$. In both cases:

  • Over the real numbers, we can locally write $y$ as a function of $x$, and this function has a well defined first derivative. I get that $y = - \sqrt[3]{4} x^{5/3} - \frac{x^2}{3} + (\mbox{higher order terms})$.

  • The second derivative, $y''(x)$, does not exist. Intuitively, $y(x) \approx x^{5/3}$, then $y'(x) \approx x^{2/3}$ and then the second derivative blows up.

  • Over the complex numbers, for $x$ arbitrarily near $0$ but not equal to $0$, there are $3$ values of $y$ which satisfy the equation, so we can't write complex $y$ as a function of complex $x$ near $0$.

I would say that this curve has a well defined tangent, with slope $0$, over the reals, but not over the complexes. Whether or not you agree depends on exactly what definition of tangent line you are using.

share|cite|improve this answer
Sorry. As far as I can see, this is not a parametrization of $\;4x^5+y^3-x^2y^2=0$ . – Han de Bruijn Feb 3 '14 at 16:26
Sign error, thanks. Try now. – David Speyer Feb 3 '14 at 16:34
'Simplify[x^2 y^2 - (4 x^5 + y^3) /. {x -> t^3/(4 - t)^2, y -> - t^5/(4 - t)^3}]' gives 0 for me in Mathematica. – David Speyer Feb 3 '14 at 16:35

The clue of this answer is actually given by David Speyer so I shall give him the credit. What follows is for the sake of completeness and clarity. $$ x(t) = \frac{t^3}{(4-t)^2} \qquad , \qquad y(t) = -\frac{t^5}{(4-t)^3} \qquad \Longrightarrow \\ x'(t) = \frac{t^2\,(12-t)}{(4-t)^3} \qquad , \qquad y'(t) = -\frac{2\,t^4\,(10-t)}{(4-t)^4} \qquad \Longrightarrow \\ \frac{dy}{dx} = \frac{y'(t)}{x'(t)} = -\frac{2\,t^2\,(10-t)}{(12-t)(4-t)} \qquad \Longrightarrow \qquad \lim_{t\rightarrow 0} \frac{y'(t)}{x'(t)} = 0 = \frac{dy}{dx}(0,0) $$ It's not necessary to use a computer algebra system for this (though I used it as well :-).
Anyway, herewith the proof is complete.

share|cite|improve this answer

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.