Calculus II ep24: The Comparison Test (Nov 5, 2026)
Overview
The lesson reviews the integral test, then develops direct and limit comparison tests for deciding whether positive-term series converge. Examples show how to compare with geometric and p-series benchmarks, choose the inequality direction that supports a conclusion, and use dominant terms to evaluate a limit comparison.
Key takeaways
- The direct comparison test proves convergence by bounding a series above with a convergent series, and proves divergence by bounding it below with a divergent series.
- For positive-term comparisons, the useful conclusions are directional: larger converges implies smaller converges, while smaller diverges implies larger diverges.
- Geometric series and p-series provide practical benchmarks: Σr^n converges for |r| < 1, while Σ1/n^p converges for p > 1.
- In expressions such as 2/(n³ + 3n + 7), removing positive denominator terms creates an upper bound, making it possible to compare with the convergent p-series Σ2/n³.
- A positive finite limit of a_n/b_n establishes that Σa_n and Σb_n have the same convergence behavior; for (n + 4)/n⁴ compared with 1/n³, that limit is 1.
- A divergent upper bound cannot prove that a smaller series diverges; for example, 1/(n + 4) ≤ 1/n does not establish divergence of the left-hand series.
Chapters
- For a decreasing function on [1, infinity), convergence of the improper integral of f(x) implies convergence of the series of f(n), and divergence of the integral implies divergence of the series.
- The integral test answers whether a series converges; it does not determine the series' exact sum.
- Examples use an improper integral with negative powers and the integral of e^(-3x); both converge, so their corresponding series converge.
- The comparison test is introduced as the next convergence tool because the integral test requires a decreasing function.
- The method compares a given series with a larger or smaller series whose behavior is already known.
- For positive terms with a_n ≤ b_n, convergence of the larger series Σb_n guarantees convergence of the smaller series Σa_n.
- If the smaller series Σa_n diverges, the larger series Σb_n must also diverge.
- The terms must be positive: with negative terms, divergence could mean tending to negative infinity, breaking the intuitive comparison argument.
- The conclusions do not reverse: divergence of the larger series alone says nothing about the smaller series.
- A geometric series converges when |r| < 1 and diverges otherwise.
- A p-series Σ1/n^p converges when p > 1 and diverges when p ≤ 1.
- Recognizing whether n appears in an exponent or as a base helps distinguish geometric-series behavior from p-series behavior.
- The denominator n³ + 3n + 7 is larger than n³, so 2/(n³ + 3n + 7) ≤ 2/n³ for positive n.
- The benchmark Σ2/n³ converges because it is a constant multiple of a p-series with p = 3.
- Since the original terms are smaller than a convergent series, the original series converges by direct comparison.
- For Σ sin²(n)/2^n, use sin²(n) ≤ 1 to bound each term by 1/2^n.
- The benchmark Σ(1/2)^n is geometric with ratio r = 1/2, so it converges.
- The original sine series converges because it is bounded above by a convergent positive-term series.
- Removing an added positive quantity from a numerator makes a positive fraction smaller; removing one from a denominator makes it larger.
- For Σ1/(n² + 4), dropping +4 gives the upper bound 1/n², a convergent p-series with p = 2.
- For Σ1/(n + 4), the bound 1/n diverges, but that only shows a larger series diverges and cannot establish divergence of the smaller original series.
- Because cos²(n) ≥ 0, each term (1 + cos²(n))/n is at least 1/n.
- The harmonic series Σ1/n diverges, so the larger series with numerator 1 + cos²(n) also diverges.
- Squaring the cosine matters: an unsquared cosine can be negative, so the same lower-bound argument would not automatically apply.
- Students apply direct comparison to two series, including one involving ln(n) and another with an exponential term and √n in the denominator.
- The examples require choosing a benchmark and an inequality direction that can actually prove convergence or divergence.
- The discussion emphasizes that a true inequality can still be unhelpful if its benchmark's convergence behavior gives no valid conclusion.
- For the series with (1 + ln n)/n, use 1 + ln(n) ≥ 1 to bound the terms below by 1/n.
- Since Σ1/n diverges, the original logarithmic series diverges by comparison from below.
- For the series with denominator 4^n + √n, discard √n to get an upper bound of 1/4^n.
- The geometric benchmark has ratio 1/4 and converges, so the original series converges.
- Instead of proving an inequality, form the ratio a_n/b_n and evaluate its limit as n approaches infinity.
- If the limit is a finite positive constant, the two positive-term series either both converge or both diverge.
- A zero limit does not satisfy the stated criterion, so the shared-convergence conclusion cannot be applied from this test alone.
- The dominant terms in (n + 4)/n⁴ suggest comparing it with n/n⁴ = 1/n³.
- The ratio ((n + 4)/n⁴)/(1/n³) simplifies to (n + 4)/n, whose limit is 1.
- Because the limit is positive and finite, the original series shares the behavior of Σ1/n³.
- The p-series with p = 3 converges, so Σ(n + 4)/n⁴ converges as well.
Transcript excerpt
0:00 Last time we were talking about the integral test. [cough] And actually, I've more or less said all that we need to say about the integral test, but I just wanted to have some examples. I have two that I wanted you all to try at the end. Could we just try them right now? I I hope that everybody remembers. So the integral test, I will just write down again what it says. It says if f is decreasing, this only works for a decreasing function on the interval 1 to infinity, then um I will say uh if the integral 1 to infinity f of x dx, you can look back on your notes from yesterday, this is what it said, then the integral converges Sorry, if the integral converges,
0:49 then the series converges. Series f of n. And uh the other way around is also true. If the integral diverges, then the sum diverges. So it means if I show you a series and that function is decreasing, if I show you a series, you can tell me if it converges or not by instead doing the integral, which is something that we we already know how to do. This is a improper integrals. All right, so your job I will just say test the convergence of And this is what, you know, your homework problems about the integral test will all be like this.
1:37 The question is not tell me what it adds up to because that's actually too hard and the integral test does not ever tell you what it adds up to. It just tells you does it converge or not. All right, test the convergence of these two things. See what you think. You should in each case turn it into an integral,
This is the start of the transcript. Save this video free for the whole transcript, searchable and time-coded.
In this course
Related notes
Summary, takeaways, and chapters were generated by AI from the video's transcript and may contain errors. The video belongs to its creator, Prof Staecker.