L'Hopital's Rule

Why a Quotient Can Be Hard to Evaluate

A quotient may have a limit that is difficult to determine even when its numerator and denominator both approach zero. Near a point where both vanish, their rates of change can reveal how their values compare. L’Hopital’s Rule makes this idea precise: under specific hypotheses, the limit of a quotient can be found from the limit of the quotient of its derivatives.

The rule is not a general instruction to differentiate the top and bottom of any quotient. Its hypotheses matter, especially the requirement that the original quotient have an indeterminate form. We will first establish a useful consequence of the Cauchy Mean Value Theorem, then prove a finite-point version of the rule and apply it in examples. We will also prove a version for quotients whose numerator and denominator grow without bound.

Definition: A quotient has the form $0/0$ as $x\to a$ if its numerator and denominator both tend to zero. It has the form $\infty/\infty$ if the absolute values of both tend to infinity. These descriptions identify situations in which the quotient’s limit is not determined by the separate limits of its numerator and denominator.

For example, knowing that $f(x)\to0$ and $g(x)\to0$ does not determine $\lim f(x)/g(x)$: the quotient might tend to zero, a nonzero number, or fail to have a limit. L’Hopital’s Rule addresses such a quotient only when the functions also satisfy suitable differentiability and nonvanishing conditions.

A Quotient Identity from Cauchy’s Mean Value Theorem

The key link is that a quotient of function values can be represented by a quotient of derivatives at an intermediate point. The Cauchy Mean Value Theorem provides that intermediate point. We state the particular consequence needed for a zero-over-zero limit.

Proposition (Intermediate Derivative Quotient): Let $a<b$. Suppose $f$ and $g$ are continuous on $[a,b]$, differentiable on $(a,b)$, and $f(a)=g(a)=0$. If $g'(t)\ne0$ for every $t\in(a,b)$, then $g(b)\ne0$, and there is a $c\in(a,b)$ such that $$ \frac{f(b)}{g(b)}=\frac{f'(c)}{g'(c)}. $$

Proof. Apply the Cauchy Mean Value Theorem to $f$ and $g$ on $[a,b]$. There is a $c\in(a,b)$ such that

\bigl(f(b)-f(a)\bigr)g'(c) = \bigl(g(b)-g(a)\bigr)f'(c).

Because $f(a)=g(a)=0$, this becomes $f(b)g'(c)=g(b)f'(c)$. First, the Mean Value Theorem applied to $g$ gives a $d\in(a,b)$ such that

$$ g(b)-g(a)=g'(d)(b-a). $$

The right-hand side is nonzero, since $g'(d)\ne0$ and $b-a>0$. Thus $g(b)\ne0$. Also $g'(c)\ne0$ by hypothesis. Dividing the Cauchy Mean Value Theorem identity by $g(b)g'(c)$ gives the asserted equality. $\square$

This proposition gives an exact equality at an intermediate point, not merely an approximation. When $b$ approaches $a$, the point $c$ lies between them and therefore approaches $a$ as well. That observation is the central limit argument behind the finite-point rule.

L’Hopital’s Rule for a Finite Point

Here is a right-hand version. A corresponding left-hand version follows by applying the same reasoning on intervals to the left of $a$. A two-sided limit can be concluded when the one-sided limits exist and agree.

Theorem (L’Hopital’s Rule, Zero-Over-Zero Form): Suppose $f$ and $g$ are continuous on $[a,a+\delta]$ for some $\delta>0$, differentiable on $(a,a+\delta)$, and satisfy $f(a)=g(a)=0$. Suppose also that $g'(x)\ne0$ for every $x\in(a,a+\delta)$. If $$ \lim_{x\to a^+}\frac{f'(x)}{g'(x)}=L, $$ where $L$ is finite or is $+\infty$ or $-\infty$, then $$ \lim_{x\to a^+}\frac{f(x)}{g(x)}=L. $$

Proof. By the Intermediate Derivative Quotient proposition, $g(x)\ne0$ for every $x\in(a,a+\delta]$: apply the proposition on $[a,x]$. For each such $x$, it also gives a point $c_x\in(a,x)$ such that

\frac{f(x)}{g(x)}=\frac{f'(c_x)}{g'(c_x)}.

Since $a<c_x<x$, the distance from $c_x$ to $a$ is at most $x-a$. Consequently $c_x\to a^+$ as $x\to a^+$. If the derivative quotient tends to a finite $L$, its values at $c_x$ tend to $L$; the displayed equality then proves that the original quotient tends to $L$. If the derivative quotient tends to $+\infty$, then for every real $M$ it is greater than $M$ whenever its argument is sufficiently close to $a$ from the right. Since $c_x\to a^+$, the same inequality holds for the original quotient for all sufficiently small positive $x-a$. This proves divergence to $+\infty$. The argument for $-\infty$ is identical with the inequality reversed. $\square$

The conclusion concerns the limit of $f/g$, not the value of $f'(a)/g'(a)$; the derivatives at $a$ need not even exist. The rule compares derivatives at nearby interior points, where $g'$ is required to be nonzero.

Worked Example: A Root Quotient at a Finite Point

Evaluate

\lim_{x\to0^+}\frac{(1+x)^{1/2}-1}{x}.

As $x\to0^+$, the numerator tends to $1-1=0$, and the denominator tends to zero. The quotient is therefore of the $0/0$ form. On a sufficiently small interval to the right of zero, the numerator and denominator are continuous at zero and differentiable in the interior. The root derivative established earlier gives

\frac{d}{dx}\bigl((1+x)^{1/2}-1\bigr) =\frac{1}{2}(1+x)^{-1/2}, \qquad \frac{d}{dx}(x)=1.

The denominator derivative is nonzero, and the derivative quotient has limit

\lim_{x\to0^+}\frac{\frac{1}{2}(1+x)^{-1/2}}{1} =\frac{1}{2}.

L’Hopital’s Rule thus gives the requested limit as $1/2$. The positive root is defined near zero on the right because $1+x>0$ there.

Worked Example: A Polynomial Quotient with a Zero Limit

For $x\to0^+$, consider

\frac{x^4+2x^5}{x^2}.

Both numerator and denominator tend to zero. Their derivatives are $4x^3+10x^4$ and $2x$, respectively; the latter is nonzero for $x>0$. The derivative quotient simplifies to

\frac{4x^3+10x^4}{2x} =2x^2+5x^3,

which tends to zero as $x\to0^+$. The rule therefore gives a limit of zero for the original quotient. This can be checked directly:

\frac{x^4+2x^5}{x^2}=x^2+2x^3\longrightarrow0.

The direct simplification confirms the answer, while the derivative quotient illustrates how L’Hopital’s Rule handles the same indeterminate form.

The Rule When Both Functions Grow Without Bound

There is also a useful version at infinity. Unlike the finite-point case, its proof must account for the values of the functions at a fixed starting point. We do this by applying the Cauchy Mean Value Theorem on a tail interval and then dividing by the growing denominator.

Theorem (L’Hopital’s Rule, Infinity-Over-Infinity Form): Suppose $f$ and $g$ are differentiable on $(A,\infty)$, $g'(x)\ne0$ there, and $|g(x)|\to\infty$ as $x\to\infty$. If $$ \lim_{x\to\infty}\frac{f'(x)}{g'(x)}=L $$ for a finite real number $L$, then $$ \lim_{x\to\infty}\frac{f(x)}{g(x)}=L. $$

Proof. Fix $\varepsilon>0$. The derivative quotient tends to $L$, so there is a $T>A$ such that for every $t>T$,

\left|\frac{f'(t)}{g'(t)}-L\right|<\varepsilon.

For any $x>T$, apply the Cauchy Mean Value Theorem to $f$ and $g$ on $[T,x]$. Differentiability ensures continuity on this closed interval, and there is a $c\in(T,x)$ for which

f(x)-f(T) = \frac{f'(c)}{g'(c)}\bigl(g(x)-g(T)\bigr).

Subtract $L g(x)$ from $f(x)$, rearrange, and use the choice of $T$:

\begin{aligned} f(x)-L g(x) &=f(T)-L g(T) +\left(\frac{f'(c)}{g'(c)}-L\right)\bigl(g(x)-g(T)\bigr),\\ \left|\frac{f(x)}{g(x)}-L\right| &\leq \frac{|f(T)-L g(T)|}{|g(x)|} +\varepsilon\frac{|g(x)-g(T)|}{|g(x)|}. \end{aligned}

As $x\to\infty$, $|g(x)|\to\infty$, so the first term tends to zero. Also,

\frac{|g(x)-g(T)|}{|g(x)|} \leq 1+\frac{|g(T)|}{|g(x)|}\longrightarrow1.

It follows that the upper limit of $\left|f(x)/g(x)-L\right|$ is at most $\varepsilon$. Since this holds for every $\varepsilon>0$, the difference tends to zero, proving the theorem. $\square$

Worked Example: Comparing Leading Polynomial Terms

Evaluate

\lim_{x\to\infty}\frac{5x^3-2x}{2x^3+7}.

The numerator and denominator both tend to $+\infty$, and the denominator derivative $6x^2$ is nonzero for $x>A=0$. Differentiating gives the derivative quotient

\frac{15x^2-2}{6x^2} =\frac{15-2/x^2}{6}.

As $x\to\infty$, $2/x^2\to0$, so this quotient tends to $15/6=5/2$. The infinity-over-infinity version therefore gives

\lim_{x\to\infty}\frac{5x^3-2x}{2x^3+7} =\frac{5}{2}.

Here the original denominator tends in absolute value to infinity, as required. The derivatives and their quotient are defined on a tail interval, which is all the theorem needs.

Checking the Form Before Applying the Rule

The most common error is to differentiate numerator and denominator without first verifying an indeterminate form. For instance, as $x\to0^+$, the quotient $(1+x)/x$ does not have the form $0/0$: its numerator tends to $1$, not zero. Differentiating both parts would give $1/1=1$, but the original quotient actually tends to $+\infty$. The hypotheses of the finite-point theorem do not hold.

A reliable application therefore checks the following points before drawing a conclusion:

Identify the form: at a finite point, verify that numerator and denominator both tend to zero; at infinity, verify that both grow without bound in absolute value.
Check differentiability: the functions must be differentiable throughout a suitable punctured interval or tail interval.
Check the denominator derivative: require $g'(x)\ne0$ on the interval where the theorem is applied.
Evaluate the derivative quotient: establish its limit rather than assuming it exists.
Apply the conclusion to the original quotient: the theorem transfers a derivative-quotient limit only under its hypotheses.

The two versions proved here do not cover every possible use of L’Hopital’s Rule. In particular, the infinity version stated here assumes a finite limit for $f'/g'$, and other forms require additional arguments or transformations. Treating the hypotheses as part of the rule—not as optional checks—keeps the conclusion logically justified.

Check Your Understanding

Use the hypotheses and arguments in this tutorial to answer the following questions.

Why does the Intermediate Derivative Quotient proposition guarantee that $g(x)\ne0$ near $a$ in the zero-over-zero setting?
In the finite-point version, why does the intermediate point $c_x$ approach $a$ as $x\to a^+$?
What must be checked before applying L’Hopital’s Rule to a quotient at a finite point?
Why does the infinity-over-infinity proof divide by a denominator whose absolute value tends to infinity?
As $x\to0^+$, why is it invalid to differentiate the numerator and denominator of $(1+x)/x$ and conclude that the original limit is $1$?

What You'll Learn