---
title: "AP Calculus BC: Radius and Endpoint Tests"
slug: ap-calculus-radius-interval-of-convergence
category: "AP"
published_at: 2026-06-05T22:30:05.615+00:00
url: https://www.apcourses.net/blog/ap-calculus-radius-interval-of-convergence
---# AP Calculus BC: Radius and Endpoint Tests

> Separate the convergence radius from endpoint decisions, compare original series and recheck brackets after differentiation or integration.

<p>A power series' radius tells you how far from its centre absolute convergence extends. It does not usually settle the boundary points. This AP Calculus BC guide uses original examples to separate the ratio test, endpoint tests and final interval, without treating a solution checklist as a fixed scoring scheme.</p>
<h2>Radius and interval answer different questions</h2>
<p>For a real power series <code>∑ cₙ(x−a)ⁿ</code>, a is the centre. With a finite positive radius R, the series converges absolutely when <code>|x−a|&lt;R</code> and diverges when <code>|x−a|&gt;R</code>. You must investigate x=a−R and x=a+R separately. The interval includes an endpoint only if the series actually converges there.</p>
<p>These ideas are in <a href="https://apcentral.collegeboard.org/media/pdf/ap-calculus-ab-and-bc-course-and-exam-description.pdf">College Board's current Calculus BC framework, topic 10.13</a>. They are BC content, not an assumption that every AP Calculus question asks for a radius. Read the requested quantity: a radius is a number, while an interval specifies which real x-values work.</p>
<h2>What the ratio test establishes</h2>
<p>For a fixed x, call the series term uₙ and, when the ratios are defined eventually, calculate <code>L=limₙ→∞|uₙ₊₁/uₙ|</code>. If L&lt;1, the series converges absolutely. If L&gt;1, including an infinite limit, it diverges. If L=1, the test is inconclusive. A zero term can make a quotient undefined, so check the centre or other such cases directly rather than dividing by zero.</p>
<p>The <a href="https://openstax.org/books/calculus-volume-2/pages/5-6-ratio-and-root-tests">ratio test applies to numerical series too</a>. At many power-series endpoints its limit is 1, which is why another test is needed. The absence of x after substitution is not a reason to prohibit the ratio test. Also, terms tending to zero are necessary for convergence, but are not sufficient: the harmonic series is a counterexample.</p>
<h2>Original worked example: one endpoint included</h2>
<p>The examples in this guide are teaching examples, not released AP questions. Let <code>S(x)=∑ₙ₌₁^∞(x−2)ⁿ/(n·3ⁿ)</code>. For x≠2, the absolute-value ratio is <code>|x−2|·n/[3(n+1)]</code>. Its limit is <code>|x−2|/3</code>. Thus absolute convergence holds for <code>|x−2|&lt;3</code>, or <code>−1&lt;x&lt;5</code>, and R=3. At the centre x=2 every term is zero, so the series converges there directly.</p>
<p>At x=−1 the series becomes <code>∑ₙ₌₁^∞(−1)ⁿ/n</code>. The positive magnitudes 1/n decrease to zero, so the alternating series test establishes convergence. The absolute-value series <code>∑1/n</code> diverges; convergence at this endpoint is conditional.</p>
<p>At x=5 the series becomes the harmonic series <code>∑ₙ₌₁^∞1/n</code>, which diverges. Here the term limit is zero and the ratio-test limit is 1; neither establishes convergence. The p-series test with p=1 gives the needed conclusion. Include the left endpoint and exclude the right: the interval is <code>[−1,5)</code>.</p>
<h2>Change the coefficients, then test again</h2>
<p>The next original variations have the same centre 2 and radius 3. They differ in their boundary behaviour. Every sum in the table starts at n=1.</p>
<table><thead><tr><th>Power series</th><th>At x=−1</th><th>At x=5</th><th>Interval</th></tr></thead><tbody><tr><td>∑(x−2)ⁿ/(n²·3ⁿ)</td><td>∑(−1)ⁿ/n² converges absolutely.</td><td>∑1/n² converges by the p-series test, p=2.</td><td>[−1,5]</td></tr><tr><td>∑(x−2)ⁿ/3ⁿ</td><td>Terms (−1)ⁿ do not tend to zero; diverges.</td><td>Terms are 1; diverges.</td><td>(−1,5)</td></tr><tr><td>∑(−1)ⁿ(x−2)ⁿ/(n·3ⁿ)</td><td>∑1/n diverges.</td><td>∑(−1)ⁿ/n converges conditionally.</td><td>(−1,5]</td></tr></tbody></table>
<p>For the n² denominator the ratio contains n²/(n+1)², whose limit is 1. For the geometric variation it is simply |x−2|/3. In the sign-changing-coefficient variation, absolute values remove (−1)ⁿ from the ratio, so the radius remains 3. At the left endpoint the two alternating factors multiply to +1; at the right only the coefficient sign remains. A right endpoint is not automatically an all-positive series, and a nonzero centre does not determine which bracket is closed.</p>
<h2>Zero and infinite radius need different conclusions</h2>
<p>For the original series <code>∑ₙ₌₁^∞(x+1)ⁿ/n!</code>, the ratio at x≠−1 is <code>|x+1|/(n+1)</code>, which tends to zero for every fixed real x. At x=−1 all terms are zero. Hence R=∞ and the interval is <code>(−∞,∞)</code>; there are no finite boundary points to test.</p>
<p>For <code>∑ₙ₌₁^∞n!(x−4)ⁿ</code>, the ratio at x≠4 is <code>(n+1)|x−4|</code>, which tends to infinity. The terms do not tend to zero and the series diverges away from 4. At x=4 all terms are zero. Thus R=0 and the convergence set is the singleton <code>{4}</code>. The empty strict inequality |x−4|&lt;0 cannot establish this centre value; direct substitution does.</p>
<h2>After differentiation or integration, recheck endpoints</h2>
<p>Term-by-term differentiation and integration preserve the radius of a power series, as topic 10.13 states. The endpoint membership can change. For the worked S(x), differentiation inside −1&lt;x&lt;5 gives <code>S′(x)=∑ₙ₌₁^∞(x−2)ⁿ⁻¹/3ⁿ</code>. At −1 its terms are (−1)ⁿ⁻¹/3, and at 5 they are 1/3. Neither tends to zero, so this derivative series has interval <code>(−1,5)</code>.</p>
<p>Integrating S from 2 to x gives <code>∑ₙ₌₁^∞(x−2)ⁿ⁺¹/[n(n+1)3ⁿ]</code> inside that open interval. At either endpoint the absolute-value terms are <code>3/[n(n+1)]≤3/n²</code>. Comparison with a convergent p-series establishes absolute convergence at both endpoints, so the integrated series has interval <code>[−1,5]</code>. These conclusions concern the resulting series; do not claim termwise differentiation is automatically valid at an original boundary point.</p>
<h2>Use a solution sequence, not a promised point table</h2>
<p>When an interval is requested, show the ratio limit or another valid radius argument, solve the strict inequality, substitute both finite endpoints, justify their convergence or divergence, and state the interval. This is a useful way to organise reasoning. It does not assign one point to each written line or guarantee a particular AP score.</p>
<p>A historical example shows why the actual task matters: <a href="https://apcentral.collegeboard.org/media/pdf/ap26-frq-calculus-bc.pdf">2026 Calculus BC Question 6</a> supplies a convergence condition and asks about a geometric sum, a derivative series, an approximation bound and a combination of series. Its <a href="https://apcentral.collegeboard.org/media/pdf/ap26-sg-calculus-bc.pdf">published scoring guidelines</a> assess those particular parts. A universal eight-point radius-and-endpoints scheme does not describe that question. Use the current framework for content and the matching released guidelines when checking a particular practice response.</p>
<p>For your next practice task, compare two series with the same radius but different coefficients. Explain every endpoint bracket before checking an answer key. If you want help reviewing that reasoning, see the <a href="/ap-calculus-bc-tuition">AP Calculus BC tuition page</a>; a study sequence or tutoring session does not guarantee an exam result.</p>
<h2>Frequently asked questions</h2>
<details><summary>What is the difference between radius and interval of convergence?</summary><div>The radius describes distance from the centre. For a finite positive radius, it gives an open interval of absolute convergence. The full interval also includes any boundary points where direct substitution produces a convergent series.</div></details>
<details><summary>Is the ratio test forbidden at an endpoint?</summary><div>No. It applies to numerical series as well as power-series terms. If its limit is 1, it is inconclusive, so choose another suitable test. Explain the resulting series and the conditions of the test you use.</div></details>
<details><summary>Does a term limit of zero prove convergence?</summary><div>No. The harmonic series has terms tending to zero and still diverges. A nonzero term limit or failure of the terms to tend to zero proves divergence; a zero limit alone leaves the question open.</div></details>
<details><summary>Do differentiation and integration keep the same convergence interval?</summary><div>They preserve the radius, but endpoint membership can change. Recheck the new series at its finite boundary points rather than copying the original brackets. Termwise operations are justified in the open interval determined by the radius.</div></details>

