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

I am trying to solve the following problem:

Let f, g be continuous positive functions $\mathbb{R}^2 \to \mathbb{R}$: show that the system of equations

$$(1-x^2)f^2(x,y) = x^2 g^2(x,y)$$ $$(1-y^2)g^2(x,y) = y^2 f^2(x,y)$$

has 4 distinct solutions on the unit circle.

Now my thinking was to use Brouwer's fixed point theorem on the unit disc: rewrite the problem as

$x = \pm \sqrt{\frac{f^2}{f^2 + g^2}},\,y = \pm \sqrt{\frac{g^2}{f^2 + g^2}}$

Now it is clear then that both of these square roots must have modulus at most 1, and furthermore, if x and y satisfy these then $x^2 + y^2 = 1$: thus it comes down to showing that these 4 $\pm$s give us the 4 distinct solutions we require. Defining

$F_{\pm \pm} = (\pm \sqrt{\frac{f^2}{f^2 + g^2}}, \pm \sqrt{\frac{g^2}{f^2 + g^2}})$ in the obvious way, each $F_{\pm \pm}$ is a continuous map to the unit disc, so each has a fixed point: so far so good I think. (Though, do we need to worry about the case where $f^2 = g^2 = 0$?)

Now, since each of the 4 functions $F_{* *}$ maps to a distinct quadrant in the unit disc, it is almost obvious that our 4 fixed points for $F_{* *}$ from Brouwer must be distinct: however, what if they both lie on the boundary where 2 quadrants meet? For example the point (1,0) is in both the "x,y positive, positive" and "x,y positive, negative" quadrant, so if 2 of our 4 fixed points coincide here then we have obviously not solved the problem. Is there a simple way around this? Or am I overlooking something? Many thanks, Peter

share|cite|improve this question
I would suggest using the Intermediate Value Theorem instead of Brouwer's theorem. – Jim Belk Jun 1 '11 at 0:53
Hi Jim: the question follows an exercise on proving Brouwer's FPT, so I assume there is an intended method which uses the theorem. Is there any way I can use it? I feel like I'm almost at a solution, barring the one degenerate case where 4 solutions coincide. – Peter Hewkin Jun 1 '11 at 9:05
@Jim Belk: sorry, I didn't realise I was meant to use your username to notify you of my response re: the above: it isn't obvious to me how to use IVT at any rate, while I feel like I almost have a workable solution with BFPT if I can just iron out the one above issue. Please do let me know if you have any thoughts. – Peter Hewkin Jun 1 '11 at 13:59

Both of your problems are solved by noting that you assumed that $f,g$ are continuous and positive. Hence $f^2\neq 0 \neq g^2$ and also that the image of your maps $F_{\pm\pm}$ are strictly inside the relevant quadrants.

If you don't assume the strict inequality, the statement is false. Let $f(x,y) = x$ and $g = 1$. Then on the unit circle where $(1-x^2) = y^2$ your equations reduce to the one equation

$$ (1-x^2) x^2 = x^2 \iff (1-x^2 - 1)x^2 = 0 \iff - x^4 = 0 $$

which only has two distinct solutions on the unit circle.

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.