Exercises — Squeeze theorem for sequences
Level 1 — Recognition
Problem 1.1
For , which pair of walls lets you squeeze, and what is the limit?
Recall Solution 1.1
WHAT bounds the messy part. The cosine of any real number lives in : that is, . This is the only fact we know about — it never settles down, but it never escapes either.
WHY divide by . We want , so divide all three parts by . Since is a positive integer (), dividing by it never flips : Walls: , . Both .
Conclusion: same limit on both walls
Problem 1.2
True or false: if for all (but the bound fails for ), we may still conclude .
Recall Solution 1.2
TRUE. A limit is a statement about the tail of a sequence. Changing or ignoring finitely many early terms cannot change where the sequence is heading. Here the threshold index is : the squeeze holds "eventually," which is exactly what the theorem requires. See the parent's remark: the bounds only need to hold eventually.
Problem 1.3
Which of these can be squeezed to a limit as stated, and which cannot? (a) (b) (c) .
Recall Solution 1.3
- (a) Both walls . Squeeze works, .
- (b) Lower wall , upper wall . Different limits — squeeze says nothing. could hover anywhere in .
- (c) Both walls . Squeeze works, .
Level 2 — Application
Problem 2.1
Find .
Recall Solution 2.1
WHAT wiggles. flips between and forever, so .
WHY divide by . Since is a positive integer, is positive for every , so dividing keeps directions: Both walls (denominator blows up). Answer: .
Problem 2.2
Find .
Recall Solution 2.2
WHAT to bound. is unpredictable but , so , hence the numerator .
WHY these constant walls help. Divide by : Lower wall , upper wall . WHY does ? A fixed constant times a sequence going to still goes to : given any target band of half-width , once we have . (This is the product rule of Algebra of limits (sum, product, quotient).) Answer: .
Problem 2.3
Find .
Recall Solution 2.3
Split the numerator. .
Bound only the wiggly piece. , so Both walls . Answer: .
Level 3 — Analysis
Problem 3.1
Find .
Recall Solution 3.1
WHY bound each denominator. There are terms, index running to . For each such : . Bigger denominator smaller fraction, so Sum over (all terms bounded identically): Limits of walls: , and . Answer: .
(Contrast with the parent's Example 4, which had and limit — the extra factor of in the numerator there changes everything.)
Problem 3.2
Find .
Recall Solution 3.2
Squeeze the dominant term. Since , take th roots. WHY may we? The th-root map is increasing on : if then (raising the smaller number to the same positive power keeps it smaller). So the inequality directions survive: That is .
WHY . This is a standard limit: for any fixed , . Intuition: to have with huge, must sit just barely above — a number even slightly bigger than , raised to a large power, overshoots . So the upper wall , the lower wall is constant . Answer: (the largest base wins).
Problem 3.3
Show that for any fixed real , where is the floor (largest integer ).
Recall Solution 3.3
WHAT floor guarantees. By definition of floor, for any real : . Put : WHY divide by . is a positive integer, so directions are preserved: Walls: lower , upper the constant . Both . Answer: .
Level 4 — Synthesis
Problem 4.1
Find .
Recall Solution 4.1
Trap the polynomial between powers of a constant. We build a sandwich valid from an explicit threshold index .
- Lower side (holds for all ): clearly .
- Upper side: we claim for all . Check: ; for we have , so . Thus the bound holds precisely from onward.
So for all : Take th roots (increasing map, directions hold):
WHY . Standard limit (proved in the parent's Example 3 by taking logs: ). So the upper wall , lower wall constant . Since the sandwich holds for all , squeeze applies. Answer: .
Problem 4.2
Find . Refer to the figure.

Recall Solution 4.2
WHAT is dangerous. oscillates and is a positive shrinking thing. Product of "wiggle" and "shrink."
Bound the wiggle, ride the shrink. Since and (positive!), multiplying preserves directions: WHY both walls die. From the parent (Example 2), , so and . In the figure this is the yellow band collapsing onto the blue line; the pink sequence, trapped inside, is forced to as well. Answer: .
Problem 4.3
Find .
Recall Solution 4.3
Use the tame bound on sine. For : (sine never overtakes its input on ; and here because each is a small non-negative angle). So for each term Sum then divide by : Upper wall , lower wall . Answer: .
Level 5 — Mastery
Problem 5.1
A student writes: " and , so by squeeze." Explain precisely why this is wrong, and give an explicit obeying the hypotheses whose limit is not .
Recall Solution 5.1
The flaw: the lower wall is the constant , whose limit is . Squeeze requires both walls to share the same limit . Here the walls disagree ( vs ), so the theorem is silent.
Counterexample. Let (so ) and for all . Then holds, yet . The wrong "conclusion" is refuted.
Problem 5.2
Prove: if for all and , then . (This "absolute-value squeeze" is the workhorse form.)
Recall Solution 5.2
Turn the absolute value into two walls. The statement means exactly Apply squeeze. Lower wall ; upper wall . Same limit , so .
(This is why "show " is the standard route: it is squeeze in disguise, resting on the $\varepsilon$–$N$ definition.)
Problem 5.3
Find using a squeeze. (No ratio test.)
Recall Solution 5.3
Write out the product and isolate a fixed head. For : Keep the last factor and bound the middle factors by : (Each factor is .) Squeeze: . Answer: .
Problem 5.4
Give a sequence that is bounded () but cannot be assigned a limit by squeeze — and say why, connecting to convergence.
Recall Solution 5.4
Take . It sits inside the walls , but those walls converge to different numbers ( and ), so squeeze gives nothing. Deeper reason: genuinely diverges — its terms cluster at both and forever, so no squeeze could ever pin it (there is nothing to pin it to). Boundedness alone does not force convergence; you need either matching walls (squeeze) or monotonicity (see Monotone convergence theorem).
Recall the whole page
Recall Rapid review (cover the right side)
Walls for (P1.1) ::: and , both , so limit . Does pin a limit? (P1.3b) ::: No — the walls have different limits, so squeeze is silent. (P2.1) ::: , walls . (P2.2) ::: , walls and . (P2.3) ::: , walls . (P3.1) ::: , walls and . (P3.2) ::: (the largest base dominates). (P3.3) ::: , via . (P4.1) ::: , trapped between and for . (P4.2) ::: , envelope . (P4.3) ::: , upper wall . Absolute-value squeeze (P5.2) ::: with gives . (P5.3) ::: (bounded above by ). Why resists squeeze (P5.4) ::: its two walls have different limits; it diverges. One-line moral of the whole page ::: two walls with the SAME limit force the trapped sequence to that limit; different-limit walls prove nothing.
Connections
- Parent: Squeeze theorem — statement, proof, base examples.
- Limit of a sequence (epsilon-N definition) — the machinery behind every "."
- Bounded sequences — where the walls come from.
- Standard limits ( n-th roots, n!/n^n, ln n / n ) — , used above.
- Monotone convergence theorem — the other route to a limit when walls don't match.
- Squeeze theorem for functions — the continuous cousin (used for ).