AP® CALCULUS AB & BC · UNIT 1

Limits using the
Squeeze Theorem

Use valid bounds to prove limits, understand the standard trigonometric limits and practise complete squeeze arguments with worked examples.

Topic 1.8 study notes5 worked examples & 4 practice problemsUpdated 21 October 2026
Open calculus notebook with limit and integral symbols, compass, set square and pencil on an olive and cream mathematics study desk

1. The Squeeze Theorem: statement and conditions

The Squeeze Theorem, also called the Sandwich Theorem or Pinching Theorem, finds a function's limit by trapping its values between two functions whose limits agree. The central task is to establish the bounds and verify the theorem's conditions.

Suppose three real-valued functions satisfy:

g(x)≤f(x)≤h(x)

The inequality must hold throughout a suitable punctured neighborhood of the target: for some positive distance δ, it holds whenever

0<|x−a|<δ

If both bounding functions approach the same finite value,

limx→ag(x)=L,limx→ah(x)=L

then the middle function has that limit too:

limx→af(x)=L
Check the neighborhood, not just the target

The bounds must hold at every relevant nearby input, not only at a few table entries or exactly at the target. The three functions need not be defined at the target, and the middle function does not need to be continuous there.

For a right-hand or left-hand limit, verify the inequality and both bounding limits on that same side. For a two-sided conclusion, the reasoning must cover both approaches. Review Topic 1.2: limit notation and one-sided limits if needed.

2. Understand what the bounds force

Think of the middle function as the filling between two slices of bread. If the lower and upper slices approach the same height, the filling has no room to approach a different height. The bounds need not meet at the target itself; what matters is their behavior nearby.

On a graph, the middle curve stays between the bounding curves. Those curves may cross or have unrelated behavior far from the target, provided the required ordering holds in a sufficiently small neighborhood. A graph can suggest the squeeze, but the inequality supplies the justification.

Upper bound3+x2→3
Trapped function3+x2sin(1x)
Lower bound3−x2→3

For nonzero inputs, sine stays between negative one and one, and the squared multiplier is nonnegative. Therefore:

3−x2≤3+x2sin(1x)≤3+x2

As the input approaches zero, both bounds approach three. The middle function therefore approaches three even though its oscillating factor has no limit. This also shows that a squeeze can produce a nonzero limit.

3. When to use a squeeze—and when it is inconclusive

  • A bounded factor times a vanishing factor. Rapid oscillation is controlled by an amplitude shrinking to zero, as in xsin(1x) or x2cos(1x).
  • An inequality is given. Evaluate the lower and upper limits before trying to find a formula for the middle function.
  • An algebraic expression has useful bounds. Absolute values, trigonometric ranges and nonnegative squares can reveal simple bounding functions.
  • A standard trigonometric limit needs justification. Geometric inequalities lead to the sine-ratio limit in radians.

A squeeze can be useful when substitution or manipulation does not finish a problem, but it need not be the only possible method. Choose an efficient procedure whose assumptions you can verify; see Topic 1.7: selecting limit procedures.

Bounded oscillation alone is not enough

The function sin(1x) has no limit at zero. Its constant bounds approach different values. A vanishing multiplier or another stronger argument is needed to force convergence.

Unequal bounding limits mean these bounds are inconclusive. They do not prove that the middle limit fails to exist. You may need tighter bounds or additional information. Likewise, a bound known only to the right cannot establish a left-hand or two-sided limit.

4. Six steps for a complete squeeze argument

  1. 1. Identify the target

    Write the function and the limiting input, including the side or infinite direction when relevant.

  2. 2. Find simple bounds

    Use a known range or absolute-value estimate. The bounds should have limits you can evaluate.

  3. 3. Justify the inequality

    Explain why it holds at every nearby input in the required domain. State restrictions and check multiplier signs.

  4. 4. Evaluate both bounding limits

    Use continuity, limit laws or an already established result. Show both values.

  5. 5. Check that the values agree

    If they differ, the proposed squeeze does not establish the answer. Look for tighter bounds.

  6. 6. State the conclusion

    Name the Squeeze Theorem and give the limit of the middle function with the original target and side.

The important work is establishing an inequality with matching bounding limits. Merely writing “by squeeze” beside an answer leaves those conditions unverified.

5. Essential inequalities and sign rules

Trigonometric ranges

For every real argument:

−1≤sin(u)≤1 −1≤cos(u)≤1

The argument can be complicated; these bounds still hold wherever it is real and defined. This includes reciprocal arguments away from zero. Do not assign a value to sine of infinity: infinity is not a real angle.

The sine-ratio inequality in radians

cos(x)≤sin(x)x≤1

This inequality applies when 0<|x|<π2. Its two-sided validity follows from the evenness of cosine and of the sine ratio.

Multipliers and absolute values

Multiplication by a positive number preserves inequalities; multiplication by a negative number reverses their direction. A squared real input is nonnegative, so it preserves the ordering. To handle both signs of a linear or odd-power input at once, use absolute values:

|xsin(1x)|≤|x| −|x|≤xsin(1x)≤|x|

In particular, the bounds with a negative and positive input as written, without absolute values, cannot be used unchanged on both sides of zero.

6. Prove the standard sine limit

limx→0sin(x)x=1

Step 1: obtain the geometric inequality

In the unit circle, compare the areas of an inner triangle, a sector and an outer tangent triangle for a positive angle below a right angle. With the angle measured in radians, this gives:

sin(x)≤x≤tan(x)

Both the input and cosine are positive in this interval. Dividing the first inequality by the input gives the upper ratio bound. Rearranging the second gives the lower cosine bound:

cos(x)≤sin(x)x≤1

Step 2: include the negative side

Cosine is even, and the quotient of sine by its input is also even: changing both signs leaves the ratio unchanged. Thus the same inequality works for nonzero angles sufficiently close to zero on either side.

Step 3: evaluate both bounds and conclude

limx→0cos(x)=1 limx→01=1

The two-sided bounds have the same limit, so the Squeeze Theorem gives the sine-ratio limit above.

The ratio is not identically one

The theorem says the ratio approaches one as the angle approaches zero. It does not say that the ratio equals one at every nonzero input, and the original ratio is undefined at zero. The value one assumes radians, not degrees.

7. Prove the cosine-difference limit

limx→01−cos(x)x=0

A direct squeeze using an absolute-value bound

The half-angle identity gives:

1−cos(x)=2sin(x2)2

For small angles, the magnitude of sine is at most the magnitude of its argument. For nonzero inputs sufficiently close to zero, apply that estimate at half the input:

|1−cos(x)x|=2sin(x2)2|x|≤2|x2|2|x|=|x|2

Consequently the quotient lies between the negative and positive versions of that vanishing bound:

−|x|2≤1−cos(x)x≤|x|2

Both bounds approach zero, so the Squeeze Theorem proves the limit. Absolute values handle the negative side correctly; the quotient itself changes sign with the denominator.

An alternative using the established sine limit

The same half-angle identity can also be rewritten, for nonzero inputs, as:

1−cos(x)x=sin(x2)x2⋅sin(x2)

The first factor approaches one and the second approaches zero, so the product approaches zero by limit laws. This is a valid alternative after establishing the sine limit; it is a product-law derivation rather than a separate direct squeeze.

8. Five worked examples with full reasoning

Example 1: a linear amplitude controls oscillation

Find limx→0xsin(1x).

Diagnosis: the reciprocal angle is undefined at zero and its sine has no limit there, so direct substitution is unavailable. Treat the sine factor through its range instead of writing sine of infinity.

−1≤sin(1x)≤1

On the positive side, multiplication by the input preserves the order. On the negative side it reverses the order:

−x≤xsin(1x)≤x,x>0 x≤xsin(1x)≤−x,x<0

Combine the two cases with absolute values:

−|x|≤xsin(1x)≤|x| limx→0(−|x|)=0,limx→0|x|=0 limx→0xsin(1x)=0

The value of the original function at zero is not needed. Defining that value to be zero would make this particular function continuous there.

Example 2: a nonnegative squared amplitude

Find limx→0x2cos(1x).

−1≤cos(1x)≤1

The square is nonnegative for all real inputs, so multiplication does not reverse the order:

−x2≤x2cos(1x)≤x2 limx→0(−x2)=0,limx→0x2=0 limx→0x2cos(1x)=0

Why it works: the amplitude tends to zero even though the cosine factor keeps oscillating. A separate limit of that factor is unnecessary.

Example 3: unequal bounds give insufficient information

Suppose the following inequality holds for positive inputs:

4x−9≤f(x)≤x2−4x+7,x>0

What can be concluded about the limit at zero? Because the assumption concerns only positive inputs, begin with a right-hand investigation:

limx→0+(4x−9)=−9 limx→0+(x2−4x+7)=7

These values differ, so these bounds do not determine the right-hand limit. They also do not show that it fails to exist. For example, both the constant zero function and the constant one function satisfy the inequality near zero on the right, yet have different limits.

Conclusion: more information or tighter bounds is needed. The left-hand behavior is not supplied, so a two-sided conclusion is not justified either.

Example 4: use the sine limit with a constant multiplier

Find limx→0sin(3x)5x, using radians.

sin(3x)5x=35⋅sin(3x)3x

The angle in the sine and the denominator now match, and that angle tends to zero. The matching sine ratio therefore tends to one.

limx→0sin(3x)5x=35⋅1=35

This uses the sine limit already justified by squeeze, together with the constant-multiple law. It does not replace the nonzero-input ratio by an exact constant.

Example 5: a squeeze at positive infinity

Find limx→∞sin(x)x.

Sine is bounded between negative one and one. For positive inputs, divide that inequality by the input without reversing its direction:

−1x≤sin(x)x≤1x limx→∞(−1x)=0,limx→∞1x=0 limx→∞sin(x)x=0

Compare the targets: this ratio tends to one at zero and zero at positive infinity. The expression alone does not identify its limit; the target matters. For negative infinity, use the absolute-value estimate or reverse the division inequalities appropriately.

9. Common squeeze patterns

Patterns, inequalities and the conditions that make them valid
Function or situationValid bounds or identityConclusion and condition
xsin(1x)−|x|≤xsin(1x)≤|x|Limit zero at zero; nonzero nearby inputs.
xnsin(1x)−|x|n≤xnsin(1x)≤|x|nLimit zero at zero for a positive integer power.
x2cos(1x)−x2≤x2cos(1x)≤x2Limit zero at zero.
sin(x)xcos(x)≤sin(x)x≤1Limit one at zero; radians and sufficiently small nonzero angles.
sin(x)x at positive infinity−1x≤sin(x)x≤1xLimit zero at positive infinity; positive inputs.
A bounded factor times a vanishing factor|A(x)B(x)| is at most M|A(x)|Limit zero if the magnitude of B is bounded by a fixed constant M nearby and A tends to zero.

For the power pattern, specifying a positive integer avoids real-domain problems with arbitrary powers at negative inputs. More generally, any real-valued amplitude approaching zero can be used if the other factor is bounded in the relevant neighborhood.

The bounded factor need not have its own limit. The ordinary product law requires appropriate component limits; the squeeze argument succeeds here because it controls the product directly.

10. Find good bounds and avoid common mistakes

A practical bounding strategy

  1. Separate the bounded part from the changing amplitude.
  2. Write a known range for the bounded part.
  3. Multiply carefully, or work with the absolute value of the whole product.
  4. Simplify to lower and upper functions whose limits are known.
  5. Check the domain, sides and matching limits before concluding.

When oscillation is present, ask whether its amplitude shrinks. The function sin(1x) alone has no limit at zero, whereas its product with the input does. Failure of direct substitution does not by itself guarantee a useful squeeze.

Common errors

  • Using unequal bounding limits. The theorem's conclusion requires equal bounding limits; otherwise this particular squeeze is inconclusive.
  • Failing to reverse inequalities for a negative multiplier. State a sign condition or use absolute values.
  • Using an odd power as though it were always positive. Absolute-value powers give consistent two-sided bounds.
  • Checking a few points instead of a neighborhood. A finite table cannot establish the required inequality for every nearby input.
  • Confusing the target value with the limit. A function may be undefined at the target and still have a squeezed limit.
  • Writing a limit statement as a pointwise identity. The sine ratio approaches one near zero; it is not identically one.
  • Ignoring units or direction. Standard trigonometric limits assume radians. Bounds at positive infinity may need changes at negative infinity.
  • Assuming a failed squeeze proves nonexistence. Different bounds or another theorem may settle the problem.

For a continuity question, finding the limit is only part of the argument. Compare it with the defined target value, using the three conditions for continuity.

11. Useful trigonometric variations

All limits in this section use radians. These identities and limit laws extend the two foundational results; check that denominators remain defined at nearby inputs.

A constant inside sine

limx→0sin(kx)x=k

For a nonzero constant, multiply a matching sine ratio by the constant. For a zero constant the original numerator is zero at every nonzero input, so the result is zero directly.

limx→0sin(kx)kx=1,k≠0

The nonzero restriction is essential in the second formula: a zero constant would make its denominator zero at every input.

The tangent ratio

tan(x)x=sin(x)x⋅1cos(x) limx→0tan(x)x=1

Cosine is nonzero near zero and approaches one, so the quotient and product laws apply.

A squared denominator changes the cosine limit

1−cos(x)x2=sin(x)x21+cos(x) limx→01−cos(x)x2=12

This identity holds at sufficiently small nonzero inputs; its extra denominator approaches two. Compare it with the original cosine-difference quotient, whose denominator has only the first power and whose limit is zero. Keep that power visible in every step.

12. Write a clear AP Calculus justification

A complete solution identifies an inequality valid near the target, evaluates both bounding limits and explains why the Squeeze Theorem applies. Use the information provided in the question, including any stated domain, graph labels or one-sided restriction.

  1. State the starting bound. For an oscillating sine or cosine factor, explain its range.
  2. Show the manipulation. State the sign of any multiplier or use an absolute-value estimate.
  3. Evaluate both limits. Do not leave the equality of the bounding values implicit.
  4. Cite the theorem and conclude. Include the original function, target and side in the final limit statement.

A reusable response structure

“The function lies between the following two bounds for all relevant inputs sufficiently near the target. Both bounding functions approach the same value. Therefore, by the Squeeze Theorem, the middle function approaches that value.” Replace those descriptions with the actual inequality and calculations.

For a multiple-choice question, first check whether the given bounds meet at the target. If they do not, distinguish cannot determine from these bounds from does not exist. For a written solution, explain the gap in the hypotheses rather than claiming the function has no limit.

If a problem supplies a graph or a table, distinguish evidence from proof: a stated valid bounding relationship is stronger than an apparent pattern at sampled points. Use graphical limit reading and table-based estimates to support interpretation without overstating what those samples show.

13. Quick reference card

Squeeze Theorem recognition and conditions
You seeThinkUse or check
xsin(1x)Bounded oscillation with shrinking amplitude.−|x|≤xsin(1x)≤|x|
x2cos(1x)The multiplier is nonnegative.−x2≤x2cos(1x)≤x2
sin(x)x near zeroThe standard sine limit.limx→0sin(x)x equals one in radians.
1−cos(x)x near zeroUse the half-angle identity or the derived bound.limx→01−cos(x)x equals zero.
sin(x)x at positive infinityBound the numerator, then divide by a positive input.−1x≤sin(x)x≤1x
Two given bounding functionsEvaluate both at the stated target.Same limit plus valid ordering gives a squeeze; different limits leave the conclusion undetermined.

14. Practice problems with reasoning

Write the inequality and the two bounding limits before opening the explanation. For the trigonometric ratio, state which established limit you use.

1. An odd-power amplitude

limx→0x3sin(1x)

Check the solution and conditions

Use absolute values so the bound works on both sides:

−|x|3≤x3sin(1x)≤|x|3

Both bounding functions approach zero, so the limit is zero. Using the unswitched bounds with the raw cubic input would give an invalid ordering on the negative side.

2. A cosine quotient at infinity

limx→∞cos(x)x

Check the solution and conditions

For positive inputs, divide the cosine range by the input:

−1x≤cos(x)x≤1x

Both bounds approach zero at positive infinity, so the limit is zero.

3. A scaled sine ratio

limx→0sin(7x)3x

Check the solution and conditions

Use radians and match the sine argument with the denominator:

sin(7x)3x=73⋅sin(7x)7x

The matching ratio tends to one, giving 73.

4. Two given polynomial bounds

2x+1≤f(x)≤x2+2x+1
Assume this holds for all sufficiently small nonzero real inputs. Find limx→0f(x).

Check the solution and conditions
limx→0(2x+1)=1limx→0(x2+2x+1)=1

The ordering is valid by the stated assumption; the upper polynomial exceeds the lower one by a nonnegative square. Both bounds tend to one, so the Squeeze Theorem gives one for the middle limit.

15. Connections and next steps

The Squeeze Theorem strengthens the method-selection toolkit: it proves a limit without requiring a simple formula for the middle function or a limit for every factor.

  • Topic 1.7: choose a bounding argument when its conditions fit the expression.
  • Topic 1.9: connect inequalities with graphical, numerical and verbal limit information.
  • Continuity: a squeezed limit can identify the value needed to fill a removable discontinuity.
  • Derivatives: the sine-ratio and cosine-difference limits help justify trigonometric derivative formulas in Unit 2; chain-rule applications follow in Unit 3.
  • Function analysis: limits and amplitude bounds help describe oscillating functions, while extrema require the relevant derivative and sign analysis in later units.
  • AP Calculus BC: related comparison and squeezing ideas reappear in sequences and series in Unit 10. A bound on sequence terms alone should not be confused with a proof of convergence of their infinite sum.

After reviewing the remaining unit topics, use the Unit 1 practice test collection to check your understanding across methods.

Independent NUM8ERS study resource. AP® is a trademark of the College Board, which does not endorse this resource.

Back to top ↑