The squeeze theorem: statement for functions and sequences, building the two bounding inequalities, limits of sin n over n and x squared sin of 1 over x, the limit of sin x over x, sums and integrals trapped between bounds (new in the 2024 Extension 2 syllabus)
“Use inequalities and the squeeze theorem to prove results about limits of sequences and functions”
If near and and both tend to , then (and the same for sequences). Build an inequality that traps the awkward quantity, show both bounds tend to the same limit and name the theorem: used for , , , sums of terms, integral bounds and . New to Extension 2 in the 2024 syllabus.
Jump to a section
What this dot point is asking
The squeeze theorem is new to the Mathematics Extension 2 11-12 Syllabus (2024), first examined in the 2027 HSC. NESA lists it in Year 12 The nature of proof, under "Proof of inequalities", alongside the triangle inequality, the arithmetic and geometric mean inequality, and graphical and calculus-based proofs. You are expected to use inequalities to prove that a limit exists and has a certain value, so the theorem sits where two ideas meet: building a correct inequality, and drawing a conclusion from it.
Suppose a runner is always ahead of a slow walker and always behind a fast walker, and the slow walker and the fast walker are heading to the same finish line. The runner has no choice but to finish at that line as well. The squeeze theorem says exactly this about numbers: a quantity that is trapped between two others cannot escape the place where they meet.
The answer
Squeeze theorem for functions. Suppose for all near (except possibly at itself) and . Then .
Squeeze theorem for sequences. If for all large and and , then .
The same statement holds as and for one-sided limits.
How to write a squeeze proof
- Find a lower bound and an upper bound for the awkward quantity. Use a bound that you can prove, such as , , or a comparison of integrands.
- State the inequality chain in one line, making sure each direction is justified (watch out for multiplying by negative numbers).
- Find the limits of the two bounding quantities and show that they are equal.
- Name the theorem and conclude with the common limit.
You do not need the theorem when the quantity has an ordinary limit by the limit laws, and you may not use it when the two bounds have different limits, because then the argument proves nothing.
Typical sources of bounds
- Bounded trigonometric functions: and trap oscillating factors, giving and .
- Largest and smallest term in a sum: a sum of positive terms lies between times its smallest term and times its largest term.
- Products with factors at most one: .
- Dominant term: .
- Integral comparison: if on then .
The classic limit
For , comparing the areas of a triangle, a circular sector and a larger triangle in a unit circle gives . Dividing by and taking reciprocals,
Because as , the squeeze theorem gives from the right, and since is an even function the left limit is the same.
How exam questions use this
- "Show that " for an oscillating factor times a decaying one.
- "Prove that the sequence converges to " for a sum, a product or a root of a sum.
- "Show that and hence find " for a reduction-formula integral.
- "Hence" parts that use an inequality proved in an earlier part as one of the bounds.
A rational expression with an oscillating term
Find .
Step 1: bound the oscillating term
For all , , so for
Step 2: limits of the bounds
Dividing the top and bottom by , and .
Step 3: conclude
By the squeeze theorem the limit is .
Marker's note: one mark for the double inequality, one for the limits of the two bounds, one for the conclusion.
A factor that oscillates faster and faster
Show that .
Step 1: bound the absolute value
For , , so , which means
Step 2: conclude
Both and tend to as , so the squeeze theorem gives the limit .
Marker's note: working with the absolute value avoids splitting into the cases and .
- Two bounds with different limits
- The squeeze theorem needs the same limit on both sides. A loose upper bound that tends to proves nothing about a quantity that tends to .
- Splitting a product when one factor has no limit
- In the factor has no limit as , so the product-of-limits law cannot be used.
- Reversing inequalities incorrectly
- Multiplying or dividing by a negative number flips the signs, as in the one-sided limits of .
- Arguing "bounded over infinity"
- Statements such as are not proofs.
- A bound that holds only for some values
- The inequality chain must be true for every (or every near ) in the range that matters, not just for a few test values. Prove each bound, then test it with a small value as a check.
Write the double inequality as the very first line of your answer, then the limits of both ends on the next line, then the conclusion in one sentence that names the squeeze theorem. If a "hence" asks you to use an earlier result, that result is nearly always one of the bounds. Test your bound with a small value such as or before you rely on it; a wrong inequality direction is the commonest way to lose all the marks.
Practice questions
Original practice questions graded from foundation to exam level, each with a full worked solution. Try them before revealing the solution.
foundation2 marksUse the squeeze theorem to show that .
Show worked solution →
Bound the numerator. For every integer , . Dividing by keeps the inequalities:
Limits of the bounds. and .
By the squeeze theorem, .
Marker's note: one mark for the two-sided bound, one for the limits of the bounds and the conclusion. Writing "" earns nothing, because has no limit.
core3 marksShow that .
Show worked solution →
Bound the oscillating factor. For , . Multiply by :
Limits of the bounds. As , both and .
By the squeeze theorem the middle expression also tends to .
Marker's note: one mark for the bound on the sine, one for multiplying through (the multiplier is positive, so the inequality directions are unchanged), one for the conclusion. The limit of alone does not exist, so you may not split the product into separate limits.
core3 marksProve that .
Show worked solution →
Write the ratio as a product of fractions.
Bound it. Each factor from to is at most and positive, so removing them can only increase the product:
Conclude. The bounds and both tend to , so by the squeeze theorem .
Marker's note: one mark for the product form, one for with a reason, one for the conclusion. A spot check: at the ratio is .
core3 marksShow that .
Show worked solution →
Lower bound. Since , we have , so .
Upper bound. Since , we have , so
Combine. . As , , so the upper bound tends to and the lower bound is the constant . By the squeeze theorem the limit is .
Marker's note: one mark for each bound, one for the limit of and the conclusion.
exam4 marksLet . Show that .
Show worked solution →
The sum has terms. Each term is for . The terms decrease as increases, so the largest is and the smallest is .
Bound the sum by times the smallest and times the largest term:
Limits of the bounds. Divide the numerator and denominator by :
By the squeeze theorem .
Marker's note: one mark for noting there are terms, one for the two bounds, one for each limit of a bound (or one for both) and one for the conclusion. Numerically, , between the bounds and .
exam4 marksFor let . Show that and hence find .
Show worked solution →
Lower bound. For the integrand is non-negative, so .
Upper bound. For , , so and therefore . Integrating over ,
Conclude. , and , so by the squeeze theorem .
Marker's note: one mark for the lower bound, one for the pointwise upper bound, one for integrating it, one for the conclusion. A check: and .
exam4 marksUse the inequality (which holds for all real ) to show that .
Show worked solution →
Case . Divide the inequality by :
As the upper bound tends to , so from the right.
Case . Dividing by reverses every inequality:
As the lower bound , so from the left.
Conclusion. Both one-sided limits equal , so .
Marker's note: one mark for each one-sided squeeze and one for noting the reversal of the inequality for negative ; one for the combined conclusion. Dividing by a negative number without reversing the signs is the classic error.