Calculus II · Unit 4A · lesson

Limits of Sequences

Concept

Learning objectives

determine whether a sequence approaches a finite limit and distinguish convergence from oscillation or unbounded growth.

Limits of Sequences

Explanation

Late behavior is the whole question

A sequence converges when its terms eventually remain as close as desired to one finite number. The phrase "eventually" permits early irregularity. A thousand strange opening terms do not matter if every sufficiently late term lies near the same target. Conversely, a graph that looks flat for the first hundred terms is not proof; later terms may wander or grow.

Sequence limits inherit much of ordinary function-limit intuition, especially when $a_n=f(n)$ and $f(x)$ has a limit as $x\to\infty$ . Still, the domain is discrete, so a sequence can converge even if an interpolating function behaves badly between integers. Always make the claim about the actual terms. Common failures are persistent oscillation, unbounded magnitude, and separate subsequences approaching different values.

Bridge

From "looks close" to "must stay close"

A graph can suggest that $a_n$ approaches $L$ , but convergence makes a stronger promise. No matter how narrow a tolerance band is drawn around $L$ , all sufficiently late terms must enter the band and remain there. A few late terms inside the band are not enough if later terms escape.

The tolerance $\varepsilon$ specifies the demanded accuracy, and the index $N$ marks the stage after which that accuracy is guaranteed. The order matters: the accuracy is chosen first, and only then may the cutoff be selected.

**Convergence is eventual trapping**
After a cutoff N, all sequence points lie inside an epsilon-band around L.
Convergence is eventual trapping. Tolerance band around L with a visible cutoff N.

Definition

Formal definition

We write $a_n\to L$ if

\forall\varepsilon>0\;\exists N\in\mathbb N\;\forall n\ge N, \qquad |a_n-L|<\varepsilon.

In ordinary language

Read the quantifiers in their actual order

An accuracy demand $\varepsilon$ is chosen first. We must then produce a cutoff $N$ , possibly depending on that demand, such that every index $n\ge N$ works. The proof is not allowed to choose a different cutoff for each later term, and a long finite run of good terms is not enough.

Proof idea

Design the cutoff from the desired error

Begin with $|a_n-L|<\varepsilon$ , solve for a sufficient lower bound on $n$ , and then choose an integer $N$ at least that large. The written proof runs forward, but the design usually runs backward.

**Three ways a sequence can fail to settle**
A convergent sequence approaches one height. Divergent sequences may oscillate, grow without bound, or wander without approaching one value.
Three ways a sequence can fail to settle. Tolerance band around L with a visible cutoff N.

How to read the visual

Convergence requires one finite target. Alternation alone does not imply convergence, and a graph that keeps rising does not converge merely because its terms are individually finite.

Concept

Definition in working language

We write $a_n\to L$ or

\lim_{n\to\infty}a_n=L

when the terms become and remain arbitrarily close to $L$ . If no finite $L$ works, the sequence diverges.

Guided walkthrough

Compute a rational sequence limit

For

a_n=\frac{2n+1}{n+3},

divide numerator and denominator by $n$ :

a_n=\frac{2+1/n}{1+3/n}\longrightarrow\frac{2+0}{1+0}=2.

The dominant linear terms determine the limit.

Worked example

A complete epsilon-N proof

Prove that $a_n=(3n+1)/n\to3$ . Since

|a_n-3|=\left|3+\frac1n-3\right|=\frac1n,

it is enough to make $1/n<\varepsilon$ , or $n>1/\varepsilon$ . Given $\varepsilon>0$ , choose

N=\left\lceil\frac1\varepsilon\right\rceil+1.

Then every $n\ge N$ satisfies $|a_n-3|<\varepsilon$ . The proof does not report a pattern; it supplies a guarantee for any requested accuracy.

Common mistake

The cutoff cannot depend on the term being checked

The number $N$ may depend on $\varepsilon$ , but after it is chosen it must work for every $n\ge N$ .

Interactive checku4a-limits_of_sequences-01

Evaluate $\lim_{n\to\infty}\frac{n}{n+1}$ .

Your work stays on this device. No account or AI grader is used.

Show hint

Divide numerator and denominator by $n$ .

Attempt once to unlock the solution

Submit an answer first. The hint is available now.

Exercise

Determine the limit of $(5n^2-1)/(2n^2+7)$ .

Exercise

Explain why $(-1)^n$ diverges.

Exercise

Determine whether $\sqrt{n^2+1}-n$ converges.

Exercise

Give a sequence with infinitely many zero terms that still converges to zero.

After the explanation

Use the section idea

Reading lens

Track the integer domain, late-term behavior, monotonicity, bounds, and any recurrence before asserting a limit.

Mental model

A sequence converges when every sufficiently late term remains arbitrarily close to one finite target.

Decision

Use algebraic limits when a formula is explicit; use bounds and monotonicity when a recurrence hides the formula.

Common trap

Reading a finite plot as proof or solving a recurrence's fixed-point equation before proving convergence.

Check yourself

Can you justify both the candidate limit and why the terms must approach it?

Source & rights

Original instruction with traceable references.

BetterGrades-original; no direct adaptation declared in the verified handoff.

Reference textbooks remain rights-separated and are not published as application assets. Any direct adaptation requires separate identification and attribution.