1/23
MATH0045 Week 4 lecture slides reco2.pdf and course notes §3. Conceptual rules and conditions: divergence, absolute convergence, comparison, geometric series, ratio, root, integral and p-tests. Audited through 17 November 2026; excludes unpublished later weeks.
What is a series?
An infinite sum: Σ from n = 1 to ∞ aₙ = a₁ + a₂ + a₃ + … . A sequence lists terms; a series adds them.
What does it mean for a series to converge?
Its partial sums Sₙ = a₁ + … + aₙ tend to a finite limit as n → ∞. That limit is the sum of the series. A series that does not converge diverges.
What does the divergence test say?
If Σaₙ converges, then aₙ → 0. Therefore, if the terms do not tend to 0, the series diverges. This is a necessary condition, not a sufficient one.
Does aₙ → 0 guarantee that Σaₙ converges?
No. The harmonic series Σ1/n diverges even though 1/n → 0. If the terms tend to 0, the divergence test is inconclusive; use another test.
What is absolute convergence?
Σaₙ is absolutely convergent if Σ|aₙ| converges. Absolute convergence implies convergence of Σaₙ.
Can a series converge without being absolutely convergent?
Yes. Σ(−1)ⁿ/n converges, but Σ1/n diverges. Such a series is called conditionally convergent.
What are the two useful conclusions of the comparison test?
For 0 ≤ aₙ ≤ bₙ: if Σbₙ converges, then Σaₙ converges; if Σaₙ diverges, then Σbₙ diverges. The inequalities need only hold from some index onwards.
Why do Σ 1/[(n+1) log(n+1)] and Σ 1/[(n+1)(log(n+1))²] behave differently?
Both summands come from positive, continuous, decreasing functions on x≥1. In each improper integral use u=log(x+1), so du=dx/(x+1). The first becomes the integral of 1/u and diverges. The second becomes the integral of 1/u² and converges. The integral test gives the same conclusions for the two series.
What comparison proves divergence of a nonnegative series?
Find a smaller nonnegative series known to diverge: 0 ≤ bₙ ≤ aₙ and Σbₙ divergent. A larger divergent series gives no conclusion about Σaₙ.
What is a geometric series and when does it converge?
a + ar + ar² + … = Σ from n = 1 to ∞ arⁿ⁻¹. For a ≠ 0 it converges exactly when |r| < 1, and its sum is a/(1 − r). If a = 0, every term is 0.
How do you identify the first term in a geometric-series sum?
Use the first term actually included, then divide by 1 − r when |r| < 1. Σ from n = 0 to ∞ rⁿ has first term 1; Σ from n = 1 to ∞ rⁿ has first term r.
What happens to a geometric series when |r| ≥ 1 and a ≠ 0?
It diverges because its terms do not tend to 0. This includes r = 1 and r = −1; alternating signs alone do not establish convergence.
What does the ratio test say?
For nonzero terms, let r = lim as n → ∞ |aₙ₊₁/aₙ|, if the limit exists. If r < 1 the series converges absolutely; if r > 1 it diverges; if r = 1 the test is inconclusive.
When is the ratio test a useful first choice?
When the terms involve factorials and/or geometric factors such as cⁿ: these often cancel in aₙ₊₁/aₙ. Use the absolute value and take the limit before concluding.
Why does a ratio-test limit of 1 give no conclusion?
Both Σ1/n and Σ1/n² have ratio limit 1. The first diverges and the second converges, so another test is needed.
What does the root test say?
Let r = lim as n → ∞ |aₙ|^(1/n), if the limit exists. If r < 1 the series converges absolutely; if r > 1 it diverges; if r = 1 the test is inconclusive.
What conditions are needed for the integral test?
Choose an integer N so that aₙ=f(n) for n≥N, where f is positive, continuous and decreasing on [N,∞). The series converges exactly when the improper integral from N to ∞ of f(x) dx is finite. A finite initial segment cannot affect convergence.
Does the integral test give the exact sum of a series?
The integral decides convergence and bounds the sum. Since f decreases, f(n+1) ≤ ∫ from n to n+1 of f(x) dx ≤ f(n). Adding these inequalities sandwiches the integral between two consecutive sums of rectangle areas. The bounds differ only by endpoint terms, which gives the integral test.
When does the p-series Σ1/nᵖ converge?
It converges for p > 1 and diverges for 0 < p ≤ 1. If p ≤ 0, the terms do not tend to 0, so it also diverges.
Why does the p-series threshold occur at p = 1?
For p ≠ 1, ∫ from 1 to b x^(−p) dx = (b^(1−p) − 1)/(1 − p), finite in the limit b → ∞ only for p > 1. At p = 1, the antiderivative is ln x and the integral diverges.
Does changing finitely many terms affect convergence?
No. A finite initial sum cannot change whether the remaining infinite tail has a finite limit. It may change the sum, but not convergence or divergence.
What is a sensible order for choosing a convergence test?
First check whether aₙ → 0. Recognise geometric or p-series; use comparison with these when possible. Factorials and powers suggest the ratio test; nth powers may suggest the root test. Positive decreasing terms may allow the integral test.
How can you prove convergence of Σ sin(n² + 1)/n²?
Use |sin(n² + 1)/n²| ≤ 1/n². The p-series Σ1/n² converges, so comparison proves absolute convergence and hence convergence of the original series.
Why does Σ(2n³ + 1)/(100n³ + n² + 1) diverge?
Its nth term tends to 2/100, not 0. The divergence test immediately shows that the series diverges; there is no need to add the terms.