Ten limits by L'Hôpital's rule: 0/0 and ∞/∞, a rule applied twice, x ln x and the 0 log 0 = 0 of entropy, xˣ and (1 + 2/x)ˣ by taking logs, a difference of fractions, and a limit the rule cannot find, with worked solutions and the mistakes of using it on a limit that is not indeterminate and of reading 1^∞ as 1.
Before you start
When substitution gives 00 or ,∞∞, L'Hôpital's rule replaces the limit of a ratio with the limit of the ratio of the derivatives. It is how the convention 0log0=0 in entropy is justified, and why x2e−x and xlnx go to 0: exponentials beat powers, and powers beat logs. The work goes wrong in a few places: the rule used on a limit that is not indeterminate, the quotient rule used instead of differentiating top and bottom separately, 1∞ read as 1, a log taken and never undone, and a conclusion drawn when the rule does not apply. These ten problems cover each of them.
L'Hôpital's rule. Suppose f and g are differentiable near a (except possibly at ),a),g′(x)=0 near ,a, and g(x)f(x) has the form 00 or ±∞±∞ as .x→a. If limg′(x)f′(x) exists (or is ),±∞), then limg(x)f(x) equals it. The same holds for one-sided limits and for .a=±∞.
f′ and g′ are differentiated separately. It is not the quotient rule.
Check the form before every use. If substitution gives a number, that number is the limit and the rule does not apply.
The other indeterminate forms are rewritten first. :0⋅∞: write the product as a quotient, .fg=1/gf.:∞−∞: combine into one fraction. ,1∞,00 and :∞0: take the log, find the limit L of the log, and the answer is .eL.
If limg′f′ does not exist, the rule says nothing: the original limit may still exist (Problem 10).
0From both sides. Near 0, ,sinx≈x−6x3, so sinx1≈x1(1+6x2) and the difference is about .−6x.
Problem 10
Find .x→∞limxx+sinx. What happens if you use L'Hôpital's rule?
As :x→∞:.∞∞.Indeterminate.
The rule gives ,x→∞lim11+cosx, which does not exist.1+cosx keeps swinging between 0 and 2. The rule needs this limit to exist, so it gives no answer here, and it does not show that the original limit fails to exist.
xx+sinx=1+xsinxDivide top and bottom by x instead.
,−x1≤xsinx≤x1, so .xsinx→0.,∣sinx∣≤1, and both bounds go to 0 (squeeze).
;1; L'Hôpital's rule does not apply, because lim(1+cosx) does not existThe rule is a sufficient condition, not a test: when the ratio of derivatives has no limit, use another method.
Where this goes wrong
1. Using the rule on a limit that is not indeterminate
The rule is used so often that it becomes the first step for every fraction.
x→1limx+1x2+1A limit to find.
“It is a fraction, so differentiate top and bottom.”The habit that causes the mistake: skipping the check of the form.
x→1lim12x=2At x=1 the fraction is ,22=1, not 00 or ,∞∞, so the rule does not apply, and the limit is 1 by substitution. Check the form every time, including before each repeat application.
2. The quotient rule instead of two separate derivatives
L'Hôpital's rule involves a fraction and derivatives, and the quotient rule is the derivative of a fraction.
,x→0limxsin3x, form 00Right so far: Problem 1, step 1.
“Differentiate the fraction.”The habit that causes the mistake: the rule remembered as "take the derivative".
x→0limx23xcos3x−sin3x=0That is the limit of the derivative of ,xsin3x, the slope of its graph near 0, which is a different question. The rule differentiates the top and the bottom separately: .13cos3x→3.
3. Reading 1^∞ as 1
One to any power is one, so a base that goes to 1 seems to settle the limit.
x→∞lim(1+x2)xRight so far: Problem 8.
“The base goes to 1, and .1x=1.”The habit that causes the mistake: letting the base reach its limit while the exponent is still finite.
x→∞lim(1+x2)x=1The base is never exactly 1, and the exponent grows as fast as the base approaches 1. At x=100 the value is already .1.02100≈7.24. Take the log: ,xln(1+x2)→2, so the limit is .e2≈7.389.
4. Forgetting to undo the log
After taking a log, the limit that comes out looks like the end of the problem.
,y=xx, so lny=xlnx→0 as x→0+Right so far: Problem 7, steps 2 and 3.
“The limit is 0.”The habit that causes the mistake: reporting the limit of lny as the limit of .y.
x→0+limxx=00 is the limit of .lny. The limit of y is .e0=1. A check: ,0.010.01≈0.955, near 1 and nowhere near 0.
5. Concluding the limit does not exist
When the rule produces something with no limit, it looks like the answer is "no limit".
,x→∞limxx+sinx, form ∞∞Right so far: Problem 10.
“L'Hôpital: lim(1+cosx) does not exist.”Right so far as a fact about .1+cosx.
So x→∞limxx+sinx does not existThe rule only says that if the ratio of derivatives has a limit, the original ratio has the same one. It says nothing when the ratio of derivatives has no limit. Here .xx+sinx=1+xsinx→1.
Print this set: lhopitals-rule.pdf (problems, answers, and worked solutions on separate pages).