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 is from Apostol's Calculus, Vol. II Section 8.3 #3(a).

Prove that $S=\{(x,y,z) \mid z^2 - x^2 - y^2 -1>0\}$ is open.

The only way to prove that a set is open which has been covered so far is to prove that for an arbitrary point $\mathop a\limits^{\small \to}   \in S$ there is an open ball $B(\mathop a\limits^{\small \to}  , r) \subset S$ by finding an explicit $r$ and showing that $\mathop x\limits^{\small \to}   \in B(\mathop a\limits^{\small \to}  , r) \implies \mathop x\limits^{\small \to}   \in S$. Nothing topological (compactness, completeness, etc.) has been discussed. For some reason I just am stuck trying to figure out an $r$ which would work.

Hints > Complete answers

share|cite|improve this question
up vote 3 down vote accepted

Choose arbitrary $(a,b,c)\in S$. Assume without loss of generality that $a,b,c\geq 0$. Put $k=c^2-a^2-b^2-1$.

Then, for any $(a',b',c')\in B((a,b,c),r)$ (with $0<r<c$, $r$ to be specified later -- notice that $c>0$) we have that $$c'^2-a'^2-b'^2-1\geq (c-r)^2-(a+r)^2-(b+r)^2-1=k-2r(a+b+c)-r^2$$ Thus, we can see that if we choose $r>0$ small enough to have $r^2+2r(a+b+c)<k$ (and $r<c$), $B((a,b,c),r)$ will be contained in $S$.

share|cite|improve this answer
For the main inequality, do you use $c'^2\ge(c-r)^2$ and so on for each one? I'm trying to see how to get there from having $\sqrt{(a'-a)^2 + (b'-b)^2 + (c'-c)^2}<r$ – process91 Aug 14 '12 at 1:51
@MichaelBoratko: Yes. This is obvious: for $(a',b',c')$ in this ball, $c'$ belongs to the interval $(c-r,c+r)$, and of course $(c-r)^2$ is smaller than the square of any number in that interval (because they're all positive). Similarly for $a',b'$. – tomasz Aug 14 '12 at 2:00
Should we also have $r<a$ and $r<b$ to be sure that $a'$ and $b'$ are positive? – process91 Aug 14 '12 at 2:06
@MichaelBoratko: You can do that, but that is not necessary. If $a\geq 0$, then $(a+r)^2$ is greater than the square of any number in the interval $(a-r,a+r)$, whether or not there are negative numbers within. – tomasz Aug 14 '12 at 2:09
OK, good deal. Thank you very much! – process91 Aug 14 '12 at 2:19

Note that $|x-x'| \leq \|(x,y,z)-(x',y',z')\|$. Furthermore, we have $|x^2-(x')^2| = |x-x'| |x+x'|$. Hence if $|x|, |x'| < L$, then $|x^2-(x')^2| \leq 2 L \|(x,y,z)-(x',y',z')\|$. The same applies to $y,y', z, z'$, of course.

Pick $(x,y,z) \in S$, let $M = \max\{|x|,|y|,|z|\}$ and $r = \min \{M, \frac{1}{6M}(z^2-(1+x^2+y^2))\}$ (which is $ >0$, of course).

Suppose $\|(x,y,z)-(x',y',z')\| < r$. Then we have $|x^2-(x')^2| < 2M r$, and similarly for the other coordinates.

Finally, we have $(z')^2-(1+(x')^2+(y')^2) > z^2-2Mr - (1 +x^2+y^2) - 4Mr$. By choice of $r$, we have $z^2 - (1 +x^2+y^2) \geq 6Mr$, from which we obtain $(z')^2-(1+(x')^2+(y')^2) > 0$, hence $(x',y',z') \in S$. It follows that $S$ is open since $B((x,y,z), r) \subset S$.

share|cite|improve this answer
Oops - I had the inequality going the wrong way in my question - I will fix it – process91 Aug 14 '12 at 1:28
My repair was interrupted by dinner! – copper.hat Aug 14 '12 at 3:04

Let $f:\mathbb{R}^3 \rightarrow \mathbb{R}$ be the function $f(x, y, z) = z^2 - x^2 - y^2 - 1$. Since $f$ is continuous and $(0, \infty)$ is open in $\mathbb{R}$, $S = f^{-1}((0, \infty))$ is open.

share|cite|improve this answer
He wrote that he it not using anything topological. – copper.hat Aug 14 '12 at 1:19
I do not have the theorem to back this up yet - only what I put above. Although I suppose I could rip into the details of this in order to reach that conclusion, I doubt that is what was intended. – process91 Aug 14 '12 at 1:19
A pity the OP can't use this since it is, by far, the easiest solution, though I believe it must be done with $\,(0,\infty)\,$ – DonAntonio Aug 14 '12 at 2:33
@DonAntonio Someone edited the question. Originally it was $(-\infty, 0)$. – Makoto Kato Aug 14 '12 at 3:36
Indeed @MakotoKato . – DonAntonio Aug 14 '12 at 3:41

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.