Turn a 0/0 limit into one you can evaluate by factoring, using a conjugate, combining fractions, or applying a trig identity, and know why the rewrite is allowed.
Topic 1.6 · 6 min read
Substituting x=5 into x−5x2−25 gives 00, which is not a number. Yet for every x=5 the fraction equals x+5, so its graph is the line y=x+5 with a single point missing at (5,10), and the outputs plainly approach 10. Every method in this topic does what that one line of factoring did: rewrite the expression into a form that agrees with it everywhere except at the troublesome point, then substitute into the new form.
What 00 does and doesn't tell you
When both the numerator and the denominator approach 0, the quotient law from topic 1.5 can't be used, and the limit could turn out to be anything. Near x=0:
Expression
Simplifies to (for x=0)
Limit as x→0
xx
1
1
x3x
3
3
xx2
x
0
x2x
x1
does not exist
All four start as 00 and end in four different places, because what matters is how fast the top and bottom shrink compared with each other. That is why 00 is called an indeterminate form: it means "rewrite and look again," never "the answer is 0," "the answer is 1," or "the limit doesn't exist."
A nonzero number over something approaching 0 is a different situation. The quotient's size grows without bound, so there is no finite limit, and you decide between ∞, −∞ and neither by checking signs on each side (topic 1.14). Algebra won't change that outcome.
The reason rewriting is allowed is that a limit at a never looks at x=a itself, only at x near a.
Canceling (x−5) from x−5x2−25 is legal only for x=5, and that is exactly the condition the box asks for.
Factoring and canceling
If a polynomial p has p(a)=0, then x−a is a factor of p (the factor theorem). So when a quotient of polynomials gives 00 at x=a, both the top and the bottom contain the factor x−a, and canceling it is the whole job. The factorings you need most:
trinomials: x2+x−6=(x+3)(x−2);
difference of squares: a2−b2=(a−b)(a+b);
difference and sum of cubes: a3−b3=(a−b)(a2+ab+b2) and a3+b3=(a+b)(a2−ab+b2).
After one cancellation, substitute again. Sometimes the result is still 00, and you factor again. Sometimes only the numerator is still 0, and then the limit is 0: x2−1x2−2x+1=(x−1)(x+1)(x−1)2=x+1x−1 for x=1, which approaches 20=0 as x→1.
When the expression isn't factored yet, expand it first. With h as the variable,
h→0limh(3+h)2−9=h→0limh6h+h2=h→0lim(6+h)=6.
Limits shaped like this one become the definition of the derivative in topic 2.2, so expanding and canceling an h is worth making automatic.
Multiplying by a conjugate
A square root in a difference, as in 9+x−3, can't be factored with integer patterns. The fix is the difference of squares in reverse: (A−B)(A+B)=A2−B2 squares the root away. The expression A+B is called the conjugate of A−B. Multiply the top and bottom by it:
Leave the conjugate in the denominator unmultiplied. You want the factor x to appear on top so it can cancel, and expanding the bottom would only hide it. The same move works when the root is in the denominator; multiply by the denominator's conjugate instead.
Combining fractions
A fraction inside a fraction usually hides the cancellation. Combine the small fractions over a common denominator first:
Watch for 3−x=−(x−3): it is the same factor with the sign flipped, and missing it flips the sign of the answer.
Trig identities
For trig expressions that give 00, an identity can produce the common factor. The ones that do the most work:
sin2x+cos2x=1, so 1−cos2x=sin2x and 1−sin2x=cos2x;
sin2x=2sinxcosx;
cos2x=cos2x−sin2x, which factors as (cosx−sinx)(cosx+sinx);
tanx=cosxsinx.
For example, sinxsin2x=sinx2sinxcosx=2cosx whenever sinx=0, so x→0limsinxsin2x=2cos0=2.
Some trig limits have no factor to cancel. x→0limxsinx is 00, but no identity turns it into something you can substitute into; it takes the squeeze theorem of topic 1.8, and topic 1.7 is about recognizing which method a limit needs.
Worked examples
Worked exampleExample 1: factoring a sum of cubes
Find x→−2limx2+x−2x3+8.
Show solution
Substitution gives 4−2−2−8+8=00, so both polynomials have the factor x+2. Factor the top as a sum of cubes with a=x, b=2, and the bottom as a trinomial:
Worked exampleExample 4: a constant that makes the limit exist (exam style, no calculator)
Let f(x)=x−2x2+kx−12 for x=2, where k is a constant.
(a) For what value of k does x→2limf(x) exist? Justify your answer.
(b) For the value of k found in part (a), find x→2limf(x).
(c) Let k=1. Find x→2−limf(x) and x→2+limf(x).
Show solution
(a) The denominator approaches 0 as x→2. If the numerator approached a nonzero number, the quotient's size would grow without bound and the limit could not exist. So the numerator must approach 0:
22+2k−12=0⟹2k=8⟹k=4.
That condition is necessary but not yet a proof the limit exists; part (b) finishes the job by finding it.
(b) With k=4, the numerator has x−2 as a factor: x2+4x−12=(x−2)(x+6). For x=2, f(x)=x+6, so
x→2limf(x)=x→2lim(x+6)=8.
(c) With k=1, the numerator approaches 4+2−12=−6 while the denominator approaches 0. For x slightly less than 2, x−2 is a small negative number, and small negativenegative is a large positive number, so x→2−limf(x)=∞. For x slightly more than 2, the denominator is a small positive number, so x→2+limf(x)=−∞.
Where students lose points
On the exam
In multiple choice without a calculator, these rewrites are what you use when substitution gives 00 and you are asked for the limit's value. In free response, they can come up inside other questions: finding a constant that makes a function continuous (topic 1.13), or evaluating a limit in a definition of the derivative.
Check yourself
Question 1
The function f satisfies f(x)=x2+1 for every x=3, and f(3)=0. What is x→3limf(x)?
A
10
B
0
C
The limit does not exist.
D
It cannot be determined from the information given.
Show answer
A
The limit at 3 only uses values of f near 3, where f(x)=x2+1, so it equals x→3lim(x2+1)=10. Choice B uses the value f(3), which a limit never looks at.
Question 2
Find x→0limx+4−2x.
Show answer
4
Multiply top and bottom by the denominator's conjugate, x+4+2. The bottom becomes (x+4)−4=x, which cancels the x on top, leaving x+4+2→2+2=4.
Question 3
What is x→π/2lim1−sin2xcosx?
A
0
B
1
C
21
D
The limit does not exist.
Show answer
D
It starts as 00. Since 1−sin2x=cos2x, the expression equals cosx1, which is 01 at 2π: it goes to ∞ from the left and −∞ from the right. Rewriting doesn't always produce a number; after a cancellation, substitute again and read what you get.
Question 4
The constants a and b satisfy x→1limx−1x2+ax+b=5. Find a and b.
Show answer
a=3 and b=−4
The denominator approaches 0 and the limit is finite, so the numerator must approach 0: 1+a+b=0. Then x−1 is a factor, x2+ax+b=(x−1)(x+c), and the limit is 1+c=5, so c=4. Expanding (x−1)(x+4)=x2+3x−4 gives a=3, b=−4.