An infinite sum is a limit of finite sums

Try

Add 1/2, then 1/4, then 1/8. The terms get smaller while the running total grows. Predict the limit of each sequence separately. Confusing these two limits is one of the most common errors in interpreting an infinite sum.

Predict

If added terms approach zero, must the total settle to a finite number? Compare adding successive powers of 1/2 with adding 1/n. Both term sequences approach zero. Their accumulated behavior requires more than that necessary condition.

Experiment

Construct a two-column ledger for aₙ=1/2ⁿ.

n	Term aₙ	Partial sum Sₙ
1	1/2	1/2
2	1/4	3/4
3	1/8	7/8
4	1/16	15/16

Draw a unit-length bar and shade half the remaining unshaded portion at each step. The shaded total is Sₙ and the latest piece is aₙ. Use the finite identity Sₙ=1−1/2ⁿ to explain why the total approaches 1 while the latest piece approaches zero.

Understand

The infinite series Σₙ₌₁∞aₙ is defined through its partial sums Sₙ=a₁+⋯+aₙ. It converges to S if this sequence has the finite limit S. Divergence can mean unbounded growth, oscillation or another failure of a finite limit. The symbol infinity does not instruct ordinary arithmetic to add a final term at an infinite index.

For n≥2, aₙ=Sₙ−Sₙ₋₁. If the series converges, both partial-sum sequences approach S, so aₙ approaches zero. This is a necessary condition. Its contrapositive is the useful divergence test: if terms fail to approach zero, the series cannot converge. Reversing the implication is invalid.

The harmonic series demonstrates the failure of that converse. Group its terms after the first into blocks ending at powers of two. From index 2^(k−1)+1 through 2^k there are 2^(k−1) terms, each at least 1/2^k. Every block contributes at least 1/2. Infinitely many blocks force unbounded partial sums, even though each individual term tends to zero. A slowly flattening finite plot cannot negate this argument.

Partial-sum formulas can also reveal terms. If Sₙ=2−1/(n+1), then for n≥2 the term is 1/n−1/(n+1). The first term is S₁=3/2, because there is no prescribed S₀ from that formula unless the problem defines it consistently. Carefully checking the starting index prevents silently changing the series. Changing finitely many terms can change its sum but not whether its tail converges.

Numerical computation has practical limits. A running total may stop changing in finite precision because a term becomes too small relative to the stored total, not because the mathematical tail is zero. Conversely, slow divergence can look stable over many iterations. Exact identities, comparison bounds and convergence theorems provide the missing tail information.

Apply

In the practice bank, name the sequence whose limit is requested. Show a finite-sum calculation before taking a limit. Prove the necessary term condition using consecutive partial sums, and use the harmonic grouping argument to explain why that condition is not sufficient.

Continue

Next derive two especially useful finite-sum identities: geometric sums and telescoping sums. Their cancellation patterns make both the accumulated value and the remaining tail visible.