Sign up ×
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.

A continuous real-valued function $f: A \to B$, such that $A, B \subseteq \mathbb{R}$ where $f(0)=1$ and $\displaystyle \lim_{x \to \infty} f(x)=-1$ must have a zero within $(0,\infty)$

However, if an asymptote existed at some $x = \alpha$, where $\alpha \notin A$ such that $f(x) \to \infty$ as $x \to \alpha^{+}$ and $f(x) \to -\infty$ as $x \to \alpha^{-}$, and in the interval $(\alpha, \infty)$, the function approaches the limit $f(x) = -1$ as $x \to \infty$, isn't there a possibility that a zero does not exist within $(0,\infty)$, thus disproving the statement?

EDIT: Also, what if $\alpha \notin A$ such that $f(x) \to \infty$ as $x \to \alpha^{-}$ and $f(x) \to -\infty$ as $x \to \alpha^{+}$? Sorry for the edit again.

share|cite|improve this question
You need to tell us what $A$ and $B$ are, because if $A$ is not connected, the theorem is false as you noticed. If $A$ is connected (or at least contains the whole of $(0,\infty)$) then the result is true, can you prove it? – Najib Idrissi Mar 17 '12 at 8:48
You shouldn't change the question after you get answers. Eventually write that you want something else at the end of the answer, but do not make the ones who gave a part of their time to give some nice answers, look like they answered another question. – Beni Bogosel Mar 17 '12 at 10:56
And you could motivate why do you really want that the function doesn't have a zero? – Beni Bogosel Mar 17 '12 at 10:58
@Beni: Sorry, I intended for the edit to be a second part to the question. – amy Mar 17 '12 at 10:58

3 Answers 3

up vote 1 down vote accepted

I will assume that $A=[0,+\infty)$. From definition of limit $$ \lim\limits_{x\to+\infty} f(x)=-1\Longleftrightarrow \forall\varepsilon>0\quad\exists\delta>0\quad\forall x\quad(x>\delta\Longrightarrow|f(x)+1|<\varepsilon) $$ Take $\varepsilon=1/2$ and for respective $\delta>0$ consider point $x_0=2\delta>\delta$, Then you get $f(x_0)<-1+\varepsilon<-1/2$.

Since $f$ is continuous on $[0,+\infty)$ it is continuous on $[0,x_0]$ and moreover $f(0)=1>0$, $f(x_0)<-1/2<0$. Then by Intermediate value theorem there exist $c\in[0,x_0]\subset[0,+\infty)$ such that $$ f(c)=0. $$

On the other hand if $A\subset[0,+\infty)$ then we can construct a counterexample. Indeed consider function $$ f(x)=-\frac{x+1}{x-1} $$ defined on $A=[0,1)\cup(1,+\infty)$ It is easy to check that $$ \lim\limits_{x\to+\infty}f(x)=-1, \qquad f(0)=1 $$ But there is no $c\in[0,1)\cup(1,+\infty)$ such that $f(c)=0$. This is my counterexample.

share|cite|improve this answer

From your question it seems that $[0,+\infty)\subset A$. A continuous function $f\colon[0,+\infty)\to\mathbb{R}$ cannot have a vertical asymptote at a point $\alpha\in(0,+\infty)$; such an $\alpha$ would be a point of discontinuity.

share|cite|improve this answer
I've edited to include the domain. Does this still apply? – amy Mar 17 '12 at 9:00
The notation $f:\mathbb{R}\to\mathbb{R}$ implies that the domain is $\mathbb{R}$. So $\alpha$ can't be a real as I suspect you want it to be. – Patrick Mar 17 '12 at 9:10
@Patrick: Thanks, edited again. – amy Mar 17 '12 at 9:13
Yes, it still applies. A continuous function $f\colon A\subset\mathbb{R}\to\mathbb{R}$ cannot have a vertical asymptote at an interior point of $A$. – Julián Aguirre Mar 17 '12 at 9:46

It seems the situation you have in mind is when $f$ is defined and continuous on $(0,a)\cup(a,+\infty)$, $f(x)\to+\infty$ when $x\to a^+$ and $f(x)\to-1$ when $x\to+\infty$ (the other hypotheses being irrelevant). Then yes, there exists $x^*$ in $(a,+\infty)$ such that $f(x^*)=0$.

The proof uses the intermediate value theorem and is as follows. First, since $f(x)\to+\infty$ when $x\to a^+$, there exists $x_1$ in $(a,+\infty)$ such that $f(x_1)\geqslant1$. Second, since $f(x)\to-1$ when $x\to+\infty$, there exists $x_2$ in $(x_1,+\infty)$ such that $f(x_2)\leqslant-\frac12$. Now, $f$ is continuous on $[x_1,x_2]$ and $f(x_1)\gt0\gt f(x_2)$, hence by the usual intermediate value theorem there exists $x^*$ in $(x_1,x_2)$ such that $f(x^*)=0$.

Edit: If the hypotheses are modified (drastically) and one now assumes that $f$ is defined and continuous on $[0,a)\cup(a,+\infty)$ with $f(0)=1$, $f(x)\to+\infty$ when $x\to a^-$, $f(x)\to-\infty$ when $x\to a^+$ and $f(x)\to-1$ when $x\to+\infty$, then there is no reason to believe there exists $x^*$ in $[0,a)\cup(a,+\infty)$ such that $f(x^*)=0$. In fact, the behaviour of $f$ on $[0,a)$ and its behaviour on $(a,+\infty)$ are unrelated and, on both intervals, the sign of $f$ may be constant.

For a specific example, consider $f(x)=\frac{a}{a-x}$ for every $0\leqslant x\lt a$ and $f(x)=\frac{x}{a-x}$ for every $x\gt a$.

share|cite|improve this answer
I've edited the question again, would this still apply? – amy Mar 17 '12 at 10:38
My post answers exactly the revised version of your question. Please read. – Did Mar 17 '12 at 11:44
"First, since $f(x) \to +\infty$ when $x\to a^{+}$ there exists $x_1$ in $(a, +\infty)$ such that $f(x_1) > 1$", but in the revised version, is this necessarily true? – amy Mar 17 '12 at 11:55

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.